I've been working a lot with mini PCs in the last year, turning them into retro games consoles, Steam Machine alternatives, and generally using them for work and home applications.

There are models for all manner of budgets, from the cheap and cheerful, all the way up to powerhouse workstations, and GMKtec is one of the brands that pops up regularly with new devices.

Such was the case at IFA 2026 in Germany, where the manufacturer launched arguably its most capable mini PC yet (if you can really call it that). Crammed into a compact aluminium case, the EVO-X5 Pro is more than a convenient computer – it's a powerful agentic AI PC that can run LLMs locally and with considerably speed.

That's because it is spec'ed through the nose. You get the AMD Ryzen AI Max+ Pro 495 processor running the show, with its 16 Zen 5 CPU cores and 32 threads. Then there's the integrated Radeon 8065S GPU, featuring 40 compute units and hardware ray tracing support.

There's a dedicated AMD XDNA 2 NPU integrated too, with 55 TOPS of AI performance for local workloads.

On top of this, the PC sports 192GB of unified LPDDR5X RAM, running at 8533MT/s. Up to 160GB of that can be assigned as VRAM too, so this is a monster of a machine.

It's not really designed for gaming – it's a more enterprise-focused machine with an eye on running AI applications and hardware creativity tools. However, those specs are more than capable of playing AAA games at 1440p with very decent frame rates.

To be honest though, while GMKtec is yet to reveal full pricing, I was told that it will not be cheap during my booth visit at the show. Indeed, if you aren't as bothered by the use of high-end AI models, and are looking more for a powerful all-rounder, I'm currently testing the brand's NEO-X1 Pro compact tower which is purposely designed for that purpose and is very much a more suitable option.

As for the EVO-X5 Pro though, the units on show proved so popular at IFA the team found them missing on the last day – the dummy versions had been stolen.

Hopefully they'll turn up. But even if not, it shows that GMKtec has something of a desirable hit on its hands.

Plaud has revolutionised the voice recording market over the last few years with its range of AI note takers. The Plaud AI Note Pro is one of the best devices on the market right now, while the Plaud Note Pin S is an equally impressive wearable recorder.

With the Plaud One, however, the company have pushed the limits of wearable AI one step further. The new device looks like a fairly standard pair of earbuds with a charging case, giving you the ability to record in two different ways: either from the earbuds or the case.

There are microphones in both the earbuds and the case, so whichever method you choose, you can record. The earbuds give you hands-free access to the AI assistant as well as the ability to easily record online meetings or phone calls, while the case can record from up to 5 metres away, thanks to the four microphones on board.

Perhaps the most interesting part to these earbuds though, is that the case has a built-in eSIM, for 4G LTE connectivity across more than 80 countries. This means that the device can connect to the Plaud Agent to upload recordings for transcription and make them available for your follow-up questions, even if your phone is offline.

The Plaud Agent platform enables you to take your transcriptions and summaries to the next level by putting the findings into tasks, drafting emails, scheduling follow-up meetings and posting in Slack. It can then output the highlighted parts in email, decks, briefs or PDF form.

Use of the Plaud Agent will now be available in the form of credits, which are included in the monthly Plaud subscription plans in addition to the minutes of transcription. The exact amounts per plan are still to be confirmed.

The Plaud One is available to preorder now from the Plaud website. The Explorer Edition is priced £229 and includes £150 in Plaud Credits.

Donald Trump has introduced a ban on the import of advanced robotic devices into the US, which includes many of today's best robot vacuum cleaners. The move forms part of the Trump administration's AI Action Plan, which aims to encourage more technology manufacturing within the United States.

Whilst the original announcement was fairly broad, the FCC has now published a document clarifying exactly what the ban covers. Under the new definition, robots that can move autonomously, are equipped with sensors, connect to a network and weigh more than 4.4lb (2kg), including any docking stations, can no longer be imported into the US.

That means the ruling applies to most robot vacuum and mop combinations, robot lawn mowers and even certain transport, pool-cleaning and industrial robots.

There is some good news for anyone who already owns one of these devices. Robots that have already received FCC certification can continue to be sold, and products that are already in homes will continue to work as normal. The ban only affects the import of future products that fall under the new rules.

Whilst it's impossible to know exactly which brands will be impacted, upcoming robot vacuums from the likes of Roborock, Dreame, Ecovacs, Narwal and Xiaomi could be delayed in the US or miss out on a launch entirely if they don't receive the necessary approvals or exemptions.

It's also worth reiterating that this is US-specific legislation, so there's currently little reason for buyers in the UK or elsewhere to worry about similar restrictions. For now, the biggest impact is likely to be on the timing and availability of future robot vacuum releases in the American market.

AI is still one of those words that sparks a strong mix of excitement and hesitation, and that's largely because it's increasingly finding its way into our homes. It almost feels expected to be woven into the latest gadgets, to the point where every new product seems to need an AI feature to be considered worthwhile.

However, as someone whose job is to cover, review and test the latest smart home tech, I don't think that's what makes this area truly useful.

I've come to realise that the best smart home gadgets are often the ones that make the annoying tasks disappear. It might be controversial to say, but none of that necessarily needs AI – it just needs technology that solves a problem you actually have.

There's always going to be a temptation to focus on the flashy side of smart home technology, and whether that comes from childhood fantasies about being a spy or living on a spaceship is another question entirely. Most launches these days can certainly make it feel that way, but some of the most useful gadgets are also the least exciting.

One of the best smart plugs isn't going to be the most impressive gadget in your home, and it's certainly not going to be the most attractive, but it can automatically turn a lamp on before you arrive or switch off an appliance when you don't need it. The same goes for smart lighting. You don't need an AI-powered bulb that understands your mood if all you want is for the hallway light to turn on when you walk through the door. All you need is a simple schedule or perhaps a motion sensor to trigger it automatically.

A robot vacuum is another good example. As a smart home reviewer, my attention naturally goes to things like suction power, mopping capabilities and the quality of its navigation and cameras. AI features such as automatic carpet detection and smart object recognition are impressive, but if the robot doesn't actually clean my home to a good standard, then I don't really care how clever it is.

That's not to say AI doesn't have a role to play. In fact, a lot of the time, it's quietly ticking away in the background without us even noticing. It can help security cameras better recognise people and pets, allow voice assistants to understand more complicated requests and even help your home adapt to your routines automatically.

However, the latest AI features shouldn't be the benchmark for deciding whether a smart home product is useful. You shouldn't have to think about whether your lights will turn off or whether you've left the heating running. When smart home tech works properly, these things just happen.

Much to many people’s dismay – including my own – AI is slowly but surely becoming the norm in many households. AI features come in many forms, including the voice assistant in your smart speaker or mapping and navigation in your robot vacuum cleaner.

More recently, AI has expanded into kitchen appliances, including fridges, ovens, air fryers and coffee machines. Depending on the brand or device you’re using, this can be as simple as an appliance with app control so you can use your machine remotely, to more advanced where AI can adjust your settings to chill or cook food more accurately.

While AI appliances are still considered a luxury for many, more and more brands are introducing AI to their range of kitchen gadgets. Personally, I think AI appliances are just making us lazier, despite these features offering some value around the home.

How AI appliances add value to your home

How AI is used in appliances varies from product to product. In most cases, AI is integrated in the appliances’ hardware or software, and is designed to take care of household tasks for you so you can get on with other things.

AI laundry appliances are a great example of this, but it’s still a ‘newer’ way to use AI that’s only really been explored by the likes of LG and Samsung, two brands that are way ahead in the AI game. AI is used in washing machines and tumble dryers via sensors that weigh your laundry and its soil level before adjusting the machine’s cycle and dispensing the right amount of detergent.

Not only does this ensure you’re washing your clothes efficiently, but it keeps your energy and water usage down to help you save money in the long term. This is another handy use case for AI appliances, especially for laundry and fridges, as AI tracks your habits and makes changes in accordance with your usage.

Speaking of fridges, a smart fridge with AI features is the only appliance I’d actually consider adding to my home. Most smart fridges have built-in cameras that recognise food and monitor your fridge’s contents.

Whether it's via the fridges’ accompanying app or the fridge door’s touchscreen, the AI works to track your food’s expiration dates and will warn you when something is going off. It can have some failings – pre-made meals in tupperware are trickier to monitor than other foods – but AI smart fridges cut down food waste, give you recipe ideas, and can adjust your fridge’s cooling during peak usage.

Are AI appliances actually needed?

With a lot of kitchen appliances, there’s a certain amount of manual operation that still needs to be done, like putting food on a baking tray to put in your oven or selecting a pre-set programme on your air fryer.

Simply put, not every appliance needs a Wi-Fi connection, a voice assistant or AI features. Where I think AI has truly lost the plot is in ovens. As mentioned, there’s still some manual intervention needed to use your oven, but there are some newer models which use internal cameras to scan your food and adjust the oven’s timing and temperatures.

Unless you’re cooking something very simple, I worry about how accurate your cooking will come out. What’s more, by the time you’ve prepped your food and put it into the oven, how much more taxing is it to switch your oven on and set the temperature? It’s hardly saving you very much time, and feels like something that won’t be used very often once the initial appeal has worn off.

Many AI appliances are luxury upgrades. As you can imagine, a smart AI fridge freezer is going to set you back at least £2,000, if not more. In my opinion, if your current appliances are still doing a job well done, you don’t need an AI version, and it’ll set you back a fair amount of money anyway, despite their ongoing financial savings.

Chrome users in the UK just got a Gemini upgrade for the desktop app – following an earlier rollout in the USA.

The addition of Gemini to the browser brings artificial intelligence (AI) benefits to the user experience and interaction with the browser.

The update will also roll out to Apple's iOS Chrome app some time in August, so iPhone users will also gain benefit.

But what exactly does this Chrome update actually bring? Here's a breakdown of the 5 key benefits:

One Tab  

You can continue to use Chrome as usual, with as many tabs as you wish, but Google's "zero-tab workspace" approach means you can summon Gemini via the icon to the upper right of the browser, even ask it to peek into other tabs to bring contextualised answers – without the need to leave your current tab.

The "side panel" to the right side of the page will pop up, negating the need to toggle between tabs, which is where you can interact with Gemini directly – and see your recent query history, as relevant.

Voice Control

Google Gemini is able to operate with text prompts, but a key part of use is with voice. Now that Gemini "lives" in Chrome – via that top-right browser icon, opening the side panel once clicked – you can also summon the AI agent by voice.

Whether you want to summarise an article, compare data points from multiple tabs, query a YouTube video's content without watching, all of this is possible – again, without leaving your current tab.

Connected Apps

As Chrome is part of the suite of Google apps, the addition of Gemini can be given permissions for cross-app context, too. Whether you're using Google's Maps, Calendar or Gmail, your AI assistant can natively look into those apps for contextual answers, when prompted.

This could mean you compose an email by voice, review it and send it to the relevant party without even opening the separate Gmail tab. Or using Maps to add a location to a Calendar meeting by merely asking Gemini directly.

Imaging

Gemini also integrates its Nano Banana 2 model – used for image creation and manipulation – meaning you can request visual results directly within your current tab. Simply text prompt or use your voice for image-based results.

Personal Protections

Now that Gemini can automate a variety of tasks for you – whether sending that email, posting to social media, or even purchasing products – it's also added personal protections in a "pause and prompt" way, always asking for additional verification before finishing a task.

Samsung has just announced the launch of its new Bespoke AI Washer Dryer, and it might actually make me look forward to doing my laundry! This 2-in-1 appliance conserves space in your home, as well as energy, as the machine comes with plenty of energy-saving features and credentials.

I have a smart meter in my home, and I always notice the days where I’ve done my laundry as my energy usage goes up significantly. Laundry machines, especially tumble dryers, use a lot of electricity which results in higher monthly bills, so a machine that can help manage these costs is a must for many households.

Samsung has clearly seen this, and is addressing this problem with its new Samsung Bespoke AI Washer Dryer. The brand has adapted the washer dryer with AI Ecobubble technology, a circulation pump that reduces your energy use by quickly turning your washing detergent into bubbles.

This action allows your detergent to wash fabrics more deeply and efficiently, so you can remove stubborn stains much quicker. It also means you can wash your clothes at lower temperatures between 20° and 30°, which has been shown to use less energy while avoiding laundry ‘casualties’ like shrinking and fading.

I’ve always been a sceptic when it comes to AI appliances, but the use of AI within the Samsung Bespoke AI Washer Dryer has actually impressed me, especially in terms of fabric care. The machine has two AI-powered cycles – Air Wash and AI Wash – that treat your clothes in a more automatic way.

Starting with AI Wash, it automatically senses the weight, soil level and softness of your washing load before adjusting the amount of water and detergent needed. It continues to make adjustments throughout the cycle, including soaking, rinsing and spinning times to better suit your laundry.

Air Wash is the cycle for your bedding and delicates. It refreshes these more sensitive fabrics without using water or detergent, and instead, blows hot air onto laundry – it’s like having a dry cleaner in your home!

I got to see Samsung’s range of Bespoke AI appliances in Milan earlier this year, including its line-up of washing machines and tumble dryers. I was incredibly impressed by its detergent dispenser and automatic cycles, as it really takes the stress out of washing, while also helping you save energy.

With the costs of appliances getting more expensive, it’s always handy to find a way to cut ongoing prices, and the Samsung Bespoke AI Washer Dryer seems like a good option for this. Its initial price, though, remains a mystery at the time of writing.

Samsung has launched the Samsung Bespoke AI Washer Dryer in 8kg - 11kg options in select markets. It’s expected to roll out globally in phases.

Google has started rolling out Gemini Spark to its native Mac app, marking a big step beyond the chatbot-style AI most of us are used to.

Rather than simply responding to prompts, Spark can carry out multi-step tasks across your Mac, from organising folders to pulling information from local files and building documents in Google Workspace.

The idea is simple: give Gemini the boring jobs. Ask it to sort all the PDFs in your Downloads folder, turn a pile of invoices into a budget spreadsheet or summarise documents into a report, and it'll get on with it.

Google says Spark only accesses folders you've explicitly shared, while future updates will even let you kick off jobs remotely from your phone and have them run on your Mac while you're away from your desk. Wow.

It's also gaining connections to apps including Canva, Dropbox, Google Keep and Google Tasks, making it more useful as a day-to-day digital assistant rather than just another AI chatbot.

There's a catch, though. Gemini Spark for macOS is currently in beta and only available to Google AI Ultra subscribers in supported countries, making it a premium feature rather than one everyone can try today.

Even so, it's a glimpse of where desktop AI is heading. Instead of asking AI for answers, we'll increasingly be asking it to tick things off our to-do list. That's the AI we were hoping for, right?

Plaud has been going great guns since the launch of its first AI product less than three years ago. It has introduced a Pro version of its note taking device, and added an upgraded NotePin S wearable to the lineup most recently.

However, its ambitions aren't to stop there, as T3 has heard another wearable is on its way this year, and it could be its first device to work independently of a mobile phone.

Its current devices, including the Plaud Note Pro and NotePin S, are able to record audio without needing a phone connection, but you do need to route AI requests through an iPhone or Android device. That includes transcribing conversations and meetings, and summarising them.

However, from what we've learned about the next Plaud device so far it could be the first with its own cellular connectivity – possibly through a built-in eSIM.

That's not confirmed as yet, but we understand that Plaud is looking into it.

We're more certain on a few other features. For starters, the new wearable will work with multiple AI agents, to be more of an assistant than just a note taker. It will also sport "at least" eight hours of battery life, with a possible 12 up its sleeve.

It will be "always on", so you can just talk to it and it'll respond. And it could learn from you as you use it more, to offer increased levels of personalisation.

We don't yet know what type of device the AI wearable will be – another pin, a smartwatch, a bracelet or necklace? But it'll have more use cases than the existing Plaud products, it seems.

Incidentally, if you are interested in what Plaud is currently up to, there's a pre-Prime Day sale on its own website right now. There's up to 20% off on bundles, including a couple of devices, or a device and an annual AI Unlimited plan.

And you can try what Plaud has to offer through its free desktop app, too.

Blink has just introduced ai Tiers to its subscription service, making its plans more advanced and accurate than before. Available on its Basic and Plus plans, the new Blink ai Tiers come with AI-powered features that make it easier to check in on your property while you’re away.

As one of the more affordable smart home brands, Blink has previously offered two subscription plans – Basic and Plus – and a free plan. As you’d expect, its free plan comes with the most simple of features, including live view, motion-activated notifications and thumbnail refresh.

For its paid Basic and Plus plans, you get a step up in features which include person and vehicle detection, notification snooze, photo capture, extended live view and live view recording. They also offered cloud storage for up to 60 days, and one of its newer features, Blink Moments which puts together recorded clips in a short timeframe into one single clip for easier viewing.

But now, Blink has taken a leaf out of its sibling’s Ring’s books, and introduced its new ai Tiers which unlock AI detection and description features for your Blink security camera or video doorbell. These new AI-powered features include Video Descriptions, Single Event Alert and Unusual Event Alert.

Starting with Video Descriptions, this new Blink AI feature uses video captured by your Blink camera or doorbell to create text descriptions to explain what’s going on in each clip. Rather than scrolling through footage or having to open the Blink app, Video Descriptions delivers text so you know what’s happening at home at a quick glance.

To cut down on notifications, Single Event Alert groups multiple motion notifications within a short period of time into one alert. It does this by comparing video clips and descriptions, and identifies similar activities, so your alerts are delivered more succinctly and aren’t too repetitive. Unusual Event Alert sends alerts when it spots unusual events that don’t usually happen around your home.

Released today, Blink’s new ai Tiers come with Blink Video Descriptions already, but Single Event Alert and Unusual Event Alert are expected to follow in future updates. More AI features are expected to launch too, although we don’t know what they are just yet.

In terms of pricing, Blink remains one of the most affordable subscriptions and only has a slight increase in price from its non-ai Tiers. The Basic ai Tier is £5.49 / $6.99 a month or £34.99 / $49.99 a year for new customers. The Plus ai Tier is £14.99 / $19.99 a month or £99.99 / $149.99 for the year.

You'll need to have been living under a rock for the last 24 hours to have missed the news of Siri AI. Apple's long-awaited AI-powered personal assistant is finally ready to launch, and was showcased at the WWDC 2026 event.

Now, Google has also made an appealing change to one of its AI-powered offerings. That sees the AI Plus subscription plan become much more appealing, with the double-whammy of a price cut and a storage boost.

Here in the UK, that sees the platform drop to just £4.49 per month with 400GB of storage. Those in the US will pay US$4.99 for the same storage level.

However you spin it, that's a great deal. You're getting twice the storage for less cash – not quite half price, but not too far from it – allowing you to make even more use of the features.

For those not familiar with Google AI Plus, the plan is designed to offer a more appealing prospect than standard Gemini for those seeking more from their AI. Users will get twice as much usage from Gemini, with greater access to Gemini 3.1 Pro and Deep Research in Gemini.

You'll also enjoy more access to image, music and video generation platforms, and more access to features in NotebookLM. And, of course, there's 400GB of cloud storage on offer, too.

It's a compelling package. If you've dabbled with the base level of AI platforms like ChatGPT or Gemini, and are looking to move to the next level, this subscription will give you a decent upgrade.

The brand also offers higher levels again, with Google AI Pro offering more usage, up to 5TB of storage and other perks like YouTube Premium. For any real hardcore users, Google AI Ultra offers 20TB of storage, substantially more access to different Gemini models and early access to new innovations.

Smart Glasses may well be cutting edge tech, but they're also at the centre of a growing wave of concerns over privacy. And that's not about to get any less controversial, as new information is coming to light about a feature baked into the devices with alarming potential.

Dubbed NameTag, the software is intended to use AI to recognise faces as seen through the cameras on the glasses. The news comes from Wired, which has found details as part of a deep dive into the Meta AI app.

Perhaps most concerningly, the report suggests that parts of that code had been added from as early as January. That seems at loggerheads with quotes from Meta in April, where it suggested that any such roll out would take "a very thoughtful approach."

Now, it's worth being really clear here – the feature is not currently enabled. Instead, it's a dormant piece of code sat in the back end.

So, the real question is whether or not this is a good thing. I can certainly see the lighter side – I'm positively dreadful at remembering names, and getting a little reminder of the name of someone I've only met once or twice before would be really handy on trips and networking events.

But there's a much more sinister potential side to this, and it's really hard to ignore. Let's be frank – this technology would allow those with bad intentions (stalkers or other abusers) log faces and use them to identify strangers in public places.

The Wired report goes on to suggest that the system pulls faceprints from Meta's servers, which will almost certainly raise further alarms. Is it possible that face data from its other platforms – the brand owns Facebook, WhatsApp and Instagram – could be used in training this system?

This is certainly going to be an interesting launch...

Microsoft CEO Sataya Nadella cast doubt on the future of apps in a recent discussion saying: "We're moving from building operating systems – devices for apps – to agents."

The statement came in a discussion with Cristiano Amon, CEO of Qualcomm, about the future of computing where AI agents act on your behalf rather than everything being about the operating system and apps.

Amon leans into this, highlighting that up until now the smartphone has been the centre of the digital world, but with AI able to understand everything, the agent becomes the centre instead, able to do the things without you opening individual software.

This all hinges on a super-personalised version of an AI agent which travels between devices. The easiest reference point is to think of JARVIS, Tony Stark's AI in Marvel movies, that can be accessed basically anywhere, at any time, to do anything.

Because this is a discussion between Microsoft on the software side and Qualcomm on the hardware side, there's consideration of the implications for hardware, with Project Solara – which Microsoft just announced during its Build developer conference – being at the heart of this.

Project Solara is about developing a new agent-first software platform that works with specialised hardware to take the next step in personalised computing.

In an introductory statement from Steven Bathiche, Microsoft says: "We imagine a diverse ecosystem of agent-first devices, from small to large, from fixed to hypermobile, from personal to professional."

One of these concept designs is shown off during the discussion – a tiny AI-equipped computer that looks like a badge on a lanyard. The second is a device more like an Amazon Echo Show with a larger screen. Both are shown in the video below, but are aimed at enterprise, rather than the consumer.

This isn't the first time that we've seen AI-first devices, but it's been a slow start. The Rabbit R1 turned out to be mostly pointless, and the Humane AI Pin suffered the same fate. While some wearables (Ray-Ban Meta, Plaud) have seen some adoption, the fundamental shift here isn't so much about computing, it's about trust.

Trusting that an AI agent can do what you need quickly and easily is a point we haven't reached yet. Yes, AI is very good at transcription and summarising meeting notes, but the voice experience isn't always perfectly smooth.

You can ask Gemini to play music, but you can also open Spotify and play the exact song you want. Sometimes the delay in Gemini is longer than it takes to do the same task manually. And when it comes to more personalised applications, will you trust AI with health and financial data, or even the data about your personal life and relationships?

While Project Solara aims to build the platform for this to happen on, it's likely to be some time before users are happy to step away from the established norms to embrace a future where the AI has all the control.

Those of you who already own a Ring doorbell will know it’s a brilliant way to boost your smart home security, especially given how strong the ecosystem is. One of the things I’ve always enjoyed most is the simplicity of the app – from checking clip history to speaking to whoever’s at the door. Even after testing plenty of the best video doorbells on the market, that’s still something I stand by today.

That said, things shifted quite a bit a few months ago when Ring rolled out its Familiar Faces feature in the UK. It might initially sound like a small addition, but in practice it’s entirely changed how I use my video doorbell day to day.

After just a few days of testing it, I quickly realised how much more useful it makes my whole setup, and I couldn't wait to share it with you any longer.

So...what is Familiar Faces?

If you didn't know, Familiar Faces is an opt-in feature that lets your Ring devices recognise people who regularly appear at your door. You can build a private directory of up to 50 known faces, and when someone approaches your camera, the system can identify them and send more specific notifications based on who it is.

Over time, you can label and manage these recognitions directly in the Ring app, helping the system become more personalised to your household.

How did it improve my security setup?

From the moment I started using Familiar Faces, the biggest change was simply not needing to open the app every single time someone rang the doorbell. Instead, I’d get a notification telling me exactly who it was. If it was someone unknown – a new neighbour, for example – I could quickly label them in the app. However, if I didn’t, it wasn’t an issue either, especially for one-off visitors like delivery drivers.

It also surprised me how fast it all works. Face recognition kicks in almost instantly once someone is in frame, with no noticeable lag or processing delay. Given how smooth the Ring app already is, I probably shouldn’t have been surprised, but it still impressed me in real use. I also found myself using the Event Summary tab often, which showed me how many visits I'd had in a day, or how many from a specific person.

Another useful addition is being able to filter your video history by name. So, for example, if I wanted to check when my mum last popped by to water the plants whilst I was away, I could simply select her name and instantly see every recorded visit. The app shows the camera used, the time, and a quick preview of the clip, without me needing to manually scroll through any footage.

Anything to note?

There are a couple of things worth being aware of, although neither is a dealbreaker. First, the person at your door does need to be in clear view of the camera for a few seconds for Familiar Faces to kick in. I’ve mainly tested it on my Video Doorbell Pro, so I'm not sure how it holds up with Ring’s best security cameras as well, but it’s not instant from a quick passing glance from my own experience.

The second thing is that, occasionally, the system can create duplicate profiles for the same person. I’ve even shown up in my own directory as “unknown” when I changed my appearance slightly – like wearing a ponytail instead of my hair down or having a hood up. It’s not constant, but it does show the AI still has some way to go with consistency in recognition.

That said, neither issue really takes away from the experience. Overall, Familiar Faces has noticeably changed how I use my Ring doorbell, making it feel more intelligent and helpful day to day – and it’s easy to see why it could be a big upgrade for a lot of users.

Familiar Faces is available now to Ring device owners with a Ring Pro (£15.99/month) and Pro Intelligence plan.

Google announced changes to its subscription plans on stage at Google I/O and while the focus was on a new $100 AI Ultra plan, it's the changes coming to the Google AI Pro plan that's got users up in arms.

The details were delivered via email to subscribers, outlining that changes to the usage limit on the AI Pro plan where changing, effective immediately.

That sees "compute-based" usage limits on the plan which take into account the complexity, features and length of the chat. This limit is refreshed every 5 hours up to your weekly limit.

This plan has 4x the usage limit of the free tier, while the Ultra tier has 5x the usage limit, according to Google's subscription page.

Those usage limits also apply to Flow and Antigravity with the 1000 AI credits that used to be included no longer included in the plan - with the option to buy AI credits if needed.

You can find your own usage under settings in the Gemini app, both in the browser and in the app, under Usage Limits. There it will tell you the percentage of use that you've reached, when it resets, and where you are on your weekly limit.

It's worth remembering that the AI Pro plan is what a lot of Pixel owners are on, because new Pixel devices come with a "free trial" of the service. The regular price for the 5TB AI Pro plan costs £18.99 / $19.99 a month.

While access to Gemini models, features and storage doesn't change, it's the usage that users are reacting to.

Over on Reddit, there are plenty of users commenting that the subscription just got a lot worse. Other users comment that for video generation it's basically unusable, while some say that switching to Claude is now an easier decision.

Google's focus for I/O was on Gemini's new skills - including Gemini Spark, the 24/7 digital assistant designed to run in the background and take on a wide range of your organisational tasks.

While the demos were impressive, Google made a point of saying that it was starting as a beta for Google AI Ultra subscribers and it raises doubts over how well it will run if you're at risk of exceeding usage limits.

While little was said about subscriptions, Google was pleased to announced the $100 AI Ultra plan - and that's $100 / £79.99 a month. For hardcore AI users who are using it for business that might be viable, but as a consumer proposition, it might not be.

Google could be about to shift up its AI-powered video creation with the introduction of Gemini Omni. The new service is thought to be an extension of the existing Veo offering, with access part of Google's paid AI Pro plan.

The current Veo model has been evolving, currently sitting on version 3.1, with audio introduced to the video creation process with version 3.

The new model is expected to be called Gemini Omni, although details leaking out have suggested that it's an extension of Veo, with familiar leaker Max Weinbach sharing metadata that labels it as "VEO_MODE_OMNI".

Shared by 9to5Google, it appears that some Gemini users have had early access – shared via Reddit – although the user sharing these details has also said that it has been revoked, with Gemini reverting back to the previous Veo experience.

In the first instance, a pop-up invitation suggested "create with Gemini Omni", before going on to say: "Meet our new video model. Remix your videos, edit directly in chat, try a template, and more."

Some of the shared results are impressive too, with photo-realistic depictions showing that this AI experience has mastered things like text, while people and their movements look realistic too – if just a little too perfect.

While there are some great results from what we've seen so far, it's still a little artificial, it shows progress towards an AI video model that's more believable – if that's something you actually want.

Part and parcel of Omni appears to be usage limits, determined by the AI plan that you subscribe to. That will allow power users to create more videos, while those on cheaper tiers are still likely to have access.

This leak comes on the eve of Google making its big announcements through Google I/O, where we're expecting a full announcement to detail the advancements of Gemini, with Omni likely to be a part of that.

We're also expecting the latest on Android, with Android 17 expected to be introduced through The Android Show later today. It looks like video creation is going to play a big part in Google's AI future – after all, "omni" means "everywhere".

Google is tipped as building a Gemini AI style agent that's made to be your personal assistant.

According to sources of Business Insider – that revealed an internal document – Google is developing an AI agent codenamed Remy.

Remy is currently being tested with employees, it's claimed, from within the staff-only version of the Gemini app. It was described as a "24/7 personal agent for work, school and daily life".

The idea here is that the AI works to help you in your life in a more integrated way. So rather than being like Gemini, that you can speak to about things, this will actually help carry out actions for you.

What is Google Remy?

Google's Remy is expected to be able to carry out actions for you, monitor areas you need covered, handle complex tasks, and it will do all this while learning about your preferences over time.

Google has not publicly commented or given any information on this project, yet.

But doesn't the AI idea itself sound familiar? Yup, this appears to be Google's play to offer something similar to the popular OpenClaw free open-source AI agent.

OpenClaw is able to respond to messages for you, manage your files, carry out research and even automate tasks across your computer without you needing to do anything. So popular was this that the Nvidia CEO, Jensen Huang, called it "definitely the next Chat-GPT".

Claude Cowork is a similar offering, allowing the AI to handle your PC tasks automatically – only with less awkward complexity than OpenClaw when it comes to setup.

So, Google getting in on that action – and offering a seamlessly simply solution – makes a lot of sense.

When will Google Remy arrive?

At this point Google has not commented on Remy publicly, suggesting it's still not at the ready-to-even-tease stage.

The race to offer an AI agent is now well and truly on, so expect Google to unveil this sooner rather than later.

In the world of AI, a new robot assistant has been revealed by Samsung, dubbed Project Luna.

Samsung took to the Milan Design Week 2026 to show off this new physical AI assistant as a concept idea. This appeared briefly as part of a larger video showcasing how Samsung sees a potential future for AI in the home.

The device itself was a table-sitting, round-faced screen that can move its "head" about and change what is displayed. It communicates using beeps and boops – yes, much like R2D2.

How does Samsung Project Luna work?

The idea here is to have an AI in your home that works for everyone living there. It can base itself in the robot body that sits on the table but can also move about. Talking to it about a TV task? It's can digitally appear on the TV and continue to assist you there.

So this would essentially be like a digital butler or conductor that works with you to control all the devices in your home as you need.

The screen gives a face-like point of interaction with a pulsating orb graphic which is like its eyes and mouth – allowing for some expression of personality. That swivel head also allows it to move and look at people interacting with it.

At this stage Project Luna is very much a design concept, showing how Samsung imagines a future of living with AI could look. So, while this may be made a reality, it's not a final design at this early stage.

In an unexpected move, Google has commented on the work between Gemini and Apple, something we rarely see from this sort of partnership. They came from Google Cloud CEO Thomas Kurian during the Google Cloud Next '26 opening keynote.

The annual conference focuses on developments in the cloud space across the Google ecosystem and is currently taking place in Las Vegas.

"We’re collaborating with Apple as their preferred cloud provider to develop the next generation of Apple Foundation Models based on Gemini technology," the Google Cloud CEO said according to 9to5Mac.

"These models will help power future Apple Intelligence features, including a more personalised Siri coming later this year," he added.

It was known from the outset that Apple had turned to Google to develop these new Foundation Models, which sit behind services like Siri and Apple Intelligence.

The addition of "a more personalised Siri" along with the release date of "later this year" reinforces the timeline that we're all working to, with an updated Apple Intelligence and smarter Siri at some point in 2026.

The partnership was first announced in January 2026, with some seeing it as admission that Apple hadn't achieved what it wanted to with its AI offering. Others see it as a savvy manoeuvre, giving Apple access to a great service while AI companies battle it out for dominance.

It's expected that this enhanced version of Apple Intelligence will be part of iOS 27, which should be released in September 2026, likely making its debut on the iPhone 18 Pro models and potentially iPhone Fold. It's also expected to be part of macOS 25.

With that in mind, there's every chance that some of the new features and technologies will be shown off at WWDC, which is confirmed for 8-12 June.

It's expected that Siri will gain agentic powers from Gemini, allowing to take actions on the users' behalf and better interact with other apps and services, with Apple's assistant also offering more of a chatbot experience like ChatGPT and Gemini.

There's an expectation that Apple will refine the AI experience too, moving from slightly behind the pack, to showcasing how AI on smartphones can offer a more personalised and truly useful experience.

Certainly, having struggled to meet expectations so far, everyone will be watching the Apple and Google partnership to see what comes next.

Google is expanding the powers that its Gemini AI has in image creation, leaning on what the company calls "Personal Intelligence". The idea is that because Gemini knows all about you, it will be able to more accurately produce the image that you're looking for.

This means Nano Banana 2 and Google Photos have access to that Personal Intelligence, so rather than having to provide a long contextual prompt to get the right image, Gemini will know your preferences and likes.

The example that Google gives is asking Gemini to "design my dream house", which it will be able to create using information it already knows about you.

Personal Intelligence leans on linking Gemini with Google apps, something that enhances the AI experience because it gives access to things like calendars, so that you can just ask Google if you have any free time and to schedule new appointments.

How can Personal Intelligence improve Gemini image creation?

As you'll already know, Google Photos is a mine of information about you. You have probably sorted photos to create event albums, you've labelled important people and animals, and it also shows all the places you've been and the things you love doing.

Once connected to Gemini, this can enhance the Personal Intelligence it has about you, meaning you can create images of you and your loved ones, for example. Because Gemini is great at creating different styles, you can apply those to your whole family.

Changing images is easy, because you can just ask Gemini to make particular adaptations and if things really don't look right, you can upload an image to help Gemini get the right style.

While some of this is already available – it's possible to create images in Gemini already – this extends the skills beyond what's currently possible. It will be rolling out to users in the US first, specifically those with AI Plus, Pro and Ultra.

It will be expanding to more in the future, but Google is keen to point out that it doesn't train its model from your Google Photos so your privacy will remain protected, and you won't find others able to create images from your private images.

One jump forward in AI that could really help a lot of people has just been shown off in the form of a robot guide dog.

Researchers at Binghamton University demonstrated the robot AI powered guide dog in a video, below. The university team said this robot dog is AI-enabled so it can physically and verbally guide the visually impaired to where they need to be.

Unlike current animal guide dogs, this robot can also talk, to share surroundings with the person, as they walk along. This could potentially enable a lot more freedom and independence – as well as spontaneity – to the visually impaired.

How does the robot guide dog work?

This robot is built using ChatGPT-4, meaning the dog can be conversational. This allows the robot to go beyond what animal guide dogs can offer their human friends.

All this means the robot can describe the trip before it begins, offering possible routes and even arrival times. Then, during the journey, the robot can give real-time spoken feedback about the environment and obstacles they encounter along the way.

The system was tested by seven legally blind participants who navigated a large multi-room office environment. The feedback was positive with the participants saying they preferred this combined approach to travel over current blind dog offerings.

If and when these AI powered guide dogs become available to the public has yet to be announced. But it could change the lives of many when they finally arrive.

YouTube is rolling out a feature that will allow you to have your own avatar represent you in YouTube Shorts, with this avatar an AI creation. YouTube says that the feature is designed to "safely and securely" provide the tools to insert yourself into videos.

The feature was first hinted at alongside YouTube talking about features to combat AI slop on social platforms, with the company outlining that this will include watermarks and digital labels so that it's clear that it's AI content - which is perhaps what it means by "safely and securely".

When the feature arrives – it's currently heading to everyone "outside Europe" according to 9to5Google – it will be integrated into the YouTube app, so it'll be quick and easy to create an AI version of yourself to use in Shorts.

Google says that the creation process only has to be done once, but you will be able to go back and update your avatar as your look changes over time. You also need to be over the age of 18 to use it and have your own YouTube channel.

How to create your AI avatar in YouTube

Here are the steps you'll need to take to create your AI avatar:

• Open the YouTube app on your phone
• Tap + to create a new video and then hit the AI star in the top right-hand corner
• Hit "create video" and then "make a video with my avatar"
• You'll then be guided through a selfie and voice capture process
• Once you're happy, you can enter a prompt to create a video using your avatar

The creation of new content – remembering that YouTube doesn't seem to think this is going to lead to a proliferation of AI slop – isn't the only thing you'll be able to do. You'll also be able to add your AI avatar to other Shorts using the YouTube Create app.

The created videos can be up to 8 seconds long – Shorts can be up to 3 minutes long – but you can create a number of different AI videos and stitch them together.

It sounds like it could be a fun tool for some creators, but it isn't going to help boost the credibility of a platform where people are already calling out fake content and getting flooded by AI generated clips. If anything, this might just legitimise the process, but move it through Google's servers instead.

If you’re in the market for smart glasses, there is a very good chance you will have heard of Ray-Ban Meta and Oakley Meta glasses. There are others, but it’s Meta’s collaboration with the two eyewear brands that are most well known.

For now at least. Samsung has already confirmed it is working on smart glasses and Apple is rumoured to be working on a pair too, but they aren’t the only ones. London-based Nothing is also said to be considering a move into this market, adding to its smartphone and audio offerings.

What could Nothing's smart glasses offer?

The news comes from Mark Gurman of Bloomberg who has reported that people familiar with the matter have claimed the company is “working on artificial intelligence-enhanced smart glasses” and “planning to release the glasses during the first half of 2027”.

According to Gurman, “Nothing’s glasses – outfitted with cameras, microphones and speakers – will rely on smartphones and the cloud to handle AI processing. The idea is to help automate tasks and offer a more personalised experience.”

It’s also claimed that Nothing’s glasses will have a similar design aesthetic to its smartphones and audio products, which will no doubt be a relief to fans. Its smartphones stand out from the crowd with features like the Glyph Interface and quirky software, while the headphones deliver bold design statements.

The report comes off the back of Meta announcing new versions of its Ray-Ban glasses for those with prescriptions so its already a few steps ahead in this category. Still, there’s plenty of scope for something a little different and Nothing is definitely a brand that delivers on that front so this is one rumour we hope does become a reality.

The report also claims Nothing is working on new earbuds with AI-focused features but Gurman didn’t note a launch timeline for those.

An ever more prominent name in AI these days is Perplexity, and now it's offering a standout new service to grow its importance even further – Perplexity Health.

The company noticed that millions of people were asking the AI health related questions, daily. Lots of these were related to medications, lab results, or symptoms. It claims that this is because the service focuses on accuracy, credible sources and verifiable citations, and so is an ideal place for health queries.

However, now it's taken that advice a step further. Perplexity Health aims to offer a more personalised experience, with potentially more accurate takes on health information, using all your already stored health data wherever it may be online.

How does Perplexity Health work?

Personal health data is currently scattered – spread across apps, devices and portals. But, Perplexity Health will take lab results from one source, for example, fitness data from another, and prescriptions where it can find them to give better suggestions based on your wellbeing history.

At launch, the AI will work with Apple Health, Fitbit, Withings and other health and fitness apps and sources. Using this data, the AI will be able to track trends across time and biomarkers.

You should have all the outputs clearly laid out in a dashboard which shows you a personalised take on your health. So when you ask a question, it will be able to draw on data from all your information.

As an example, the company says: "A question about resting heart rate, for example, can factor in your recent activity, your cardiac history, and your latest blood work."

Or build a custom marathon training plan based on all your data for the most optimum, safe progress.

Perplexity Health on Computer will first become available for Pro and Max users in the US over the coming weeks, spreading to more users and locations "soon".

One of the biggest names in AI, OpenAI, has announced that it is working on a "superapp" that combines its many offerings, including ChatGPT.

The idea, confirmed this week in a Wall Street Journal report, is to combine OpenAI's apps in order to offer a more streamlined solution. This "superapp" will combine ChatGPT, the company's coding system Codex, and Altas the company's browser.

Not only will this simplify production for OpenAI and ensure greater output efficiency, it should make for a better all-in-one user experience too.

How does the OpenAI "superapp" work?

By combining the Altas browser, ChatGPT and the Codex automation, it should make for an experience that's far more seamless. This means you could do a lot more all in one place.

For example, you could ask questions about what you're browsing and have AI answers, direct. Automation of tasks right there in the browser suddenly becomes a reality too.

For coders, it means writing and debugging code right there using Codex. And, of course, with ChatGPT baked into all this, you'll constantly have an AI assistant readily available.

This shift makes sense, not only for the companies offering the apps but also for us end users.

No more need for multiple tabs open that don't link. Instead a seamless AI-shared experience could make working together easier than ever.

There's no timeline yet on when this OpenAI "superapp" might appear, but we'd expect it to be available for PCs and Macs before phones. It could then roll out to other platforms over time.

The age of robotics and AI is truly here and that's very apparent, as it's arrived in McDonalds. Yup, the major food chain has begin testing robotic staff in its actual restaurants.

The Shanghai McDonalds was the home of Keenon Robotics machines, which went to work serving customers this week.

The bots covered a host of tasks, from greeting and providing information, cleaning, to delivering food to customers, and even collecting trays.

How will robots work at McDonalds?

While this was a test, the future of robots in restaurants could look similar to this setup. The idea is to go fully automated with a single location no longer requiring many human staff at all.

From front end human interactions with service to back end kitchen staff cooking the food, McDonalds is looking into making it all robot run.

The reality is that this was very much a test and the idea that this could work as a restaurant, at this early stage, is still a reach.

While robots running restaurants is still years away, this could sign us edging that much closer to humanoid robots working alongside people in the near future.

The androids serving customers certain look capable, while the wheel based screen-toting bots appear far more fun.

What all this means for jobs, the economy and our future is a far bigger question that this burger based trial can't answer right now.

There are plenty of universal experiences in this world, but one of the biggest has to be standing in front of the fridge, staring inside absent-mindedly with absolutely no idea what to make for dinner. More often than not, it ends with ordering a last-minute takeaway or making pesto pasta for the third time that week.

It’s exactly this kind of everyday dilemma that different brands are trying to solve, with many focusing on how smart technology can act as a personal kitchen companion. After being invited to an event hosted by Siemens last week, the question has been firmly on my mind: should we be using AI in our everyday cooking, or should creating food remain a human-led process?

Interestingly, it was Siemens itself that conducted research into this very issue. The brand found that six in ten Brits struggle at least once a week to decide what to cook, with one in five facing that dilemma every single day. This is how Siemens decided on hosting its Chef’s TAIble event at its London showroom – an intimate dining experience designed to demonstrate how AI and smart appliances can work together in practice.

Before the event, guests were asked to select a few ingredients and flavour preferences. AI then generated a bespoke four-course menu in real time, which was brought to life by chef Jozef Youssef using a range of Siemens’ smart appliances.

There was something undeniably special about telling a computer what flavours you enjoy, what foods you like or dislike, and watching it produce a menu tailored entirely to your own taste buds. From a practical perspective, it makes a lot of sense. After all, coming up with something new to cook every single day can be difficult, and when an AI generator tells you what to cook, it starts to feel like a pretty logical step forward.

However, will AI recipe generators and smart chefs eventually take over our kitchen creativity? Probably not just yet… and here’s why.

Despite everything it can do, it’s easy to forget that AI is, at the end of the day, still just a computer. Things like patchy Wi-Fi, limited inputs and the tendency to repeat similar ideas are still very real limitations. In fact, some of this became clear during the event itself.

After the event, I then spoke with Jesse Kempner, Head Chef at Mindful Chef, and our conversation reminded me that not all professional chefs enjoy the influence of AI.

She explained that "In recipe development, things like balance, texture, seasonality and how ingredients behave when cooked are critical. Those nuances still come from the experience of chefs actually testing and tasting dishes in the kitchen. When I create recipes with the team, the process is about those random, spontaneous moments of creativity that only come from the joy that experimenting in the kitchen unlocks.”

That comment got me thinking. If we had simply sat down and described our flavour preferences directly to Jozef himself, and left him to create the menu without AI, would the results have been different? Most likely yes, and perhaps even a little more varied.

Recipe services like Mindful Chef are also designed to tackle the exact same problem Siemens is trying to solve – the question of what to cook. In some ways, a curated recipe box might still be the simpler solution for many people.

That said, what I find most exciting about Siemens’ approach isn’t necessarily AI telling you what to cook, but rather how to cook it. By connecting smart appliances to an AI system that can guide you through the cooking process step by step, it helps people get far more out of the appliances they already own, especially if they’ve invested a lot of money in them.

Helping people understand their appliances better and achieve better results in the kitchen feels like a much more practical use of AI to me. It’s a subtle shift, but one that makes a lot of sense, and it could quietly become one of the most useful ways AI finds its place in our kitchens.

Now that microphones are almost everywhere and near always listening, the addition of AI represents a real potential invasion of privacy.

To help protect your privacy, from prying AI surveillance, Deveillance has announced its Spectre I. This is a physical device that actively blocks microphones from listening to your private conversations.

The startup created this new type of device as a portable protector that allows you to feel secure in privacy, no matter where you are.

How does the Deveillance Spectre I work?

The Spectre I works to make conversations unintelligible to other devices. It does this by creating a two metre protection zone around the person. Activate the device and it will scan for nearby microphones and emit signals that these can detect but the human ear cannot.

The result is that your speech is overlaid with these signals, rendering the audio misunderstood by listening devices.

The device does this using a combination of signal processing, AI and some "physics based research" – which sounds a lot like the secret sauce to how this really works, so don't expect that to be shared openly by the company.

The unit itself is small and portable so could be carried to a business meeting and set down on the table, much like a speakerphone might sit centralised in the room. There is a subtle light so you know it is active and working to keep your conversations private.

The Spectre I is currently open for pre-orders where you can place a refundable deposit of $1,199/£900 to ensure you receive one first, when these begin to roll out with initial shipments expected in the second half of 2026.

One of the best AI platforms has just had an upgrade that makes it even better in the form of ChatGPT-5.3 Instant.

This new update, announced by creator OpenAI, is all about better responses to questions and requests. Previously when you asked a question, it was met with an answer that started with a wordy reflection of what you said.

This felt more like ego massaging and positive feedback, than anything that was particularly useful in terms of what you were looking for.

In fact, in an age where compute means real-wold environmental costs at the server level, this was downright wasteful. Not to mention borderline manipulative as it attempted to illicit a dopamine response in effectively feeling praised.

What's new in ChatGPT-5.3 Instant?

The latest ChatGPT-5.3 Instant update fixes this old problem by streamlining how responses work. Now the model does not begin its response with moralizing or sycophantic babble, instead getting right to the heart of the issue at hand.

Ask a question. Get an answer. That's more the vibe for this updated version of the smart AI assistant. The result? Everything should start to feel less patronising and potentially awkward.

Also added is a better blending of responses and search results. Plus, answers are now brought to the top, rather than burying them in paragraphs. These should also be more accurate as the company says hallucination rates have dropped by up to 26.8%.

All that not only means we're getting a better AI agent, but it also shows the companies behind the AI are hearing what people want and implementing that.

ChatGPT-5.3 Instant has begun rolling out and should be available to you now.

In the age of AI excitement, there is a darker underbelly as the music of artists is being cloned and effectively stolen to be re-worked by AI agents.

Now, My Music My Choice has been created as the antidote to that problem. Its creators say that it is able to stop AI from cloning music in a bid to protect artists' work and preserve copyrights, across Spotify and the like.

Just last year AI clones of Drake and Bad Bunny landed on streaming platforms, undetectable as different to listeners, but a painful hit to the music industry.

Now, thanks to researchers at Binghampton University and startup Cauth AI, that problem could be solved. My Music My Choice, or MMMC, effectively poisons the music so AI cloners can't get access to it freely.

How does My Music My Choice work?

The MMMC software works pre-emptively. Unlike current solutions to the problem of music cloning, which work after the deed is done to try and shut down fakes, this gets out ahead of the problem.

The software is able to poison the recording before it is shared publicly. This is done using microscopic alterations that leaves the music sounding the same to you and I. But put that into cloning software and it will confuse the algorithm, producing only distorted static as a result.

That means this should leave listeners having the same high quality experience, via whatever listening platform they use. But for artists it will mean their work is protected so that they receive the rewards, only, and not anyone trying to clone it and steal the potential revenue.

The other issue, aside from financial loss, is that cloning can lead to anyone using an artist's voice to say something that was not really expressed by them. That could represent far worse longer term damage to the artist.

MMMC has been tested on 150 tracks across various genres as plans to scale it up are underway. So expect artist protection from AI cloning to hit the mainstream soon.

Major trade shows are traditionally packed with launches of new devices set to launch in the coming months, but we also often get to see tech concepts that may or may not make it to market.

Lenovo is synonymous with this – often using the likes of CES and IFA to show interesting innovations that wow us. And to be fair, many of them do either become a product down the line, or form the basis of one.

We certainly hope some of the concepts shown at Mobile World Congress in Barcelona this week turn into something we can buy and use. As an owner of a Legion Go S (SteamOS) gaming handheld, I'm particularly interested in the Legion Go Fold Concept, for example.

But perhaps the most intriguing device making its debut at MWC is a desktop robot that takes the concept of smart assistants to a whole new level.

The Lenovo AI Workmate Concept is a Wall-E style 'bot that sits on your desk and aids you with productivity. It has multimodel input capture, enabling you to interact with projections it spews onto the worktop, and it's powered by AI (of course).

The AI Workmate even has a pair of goggly eyes to make it seem more friendly and useful – like a physical version of Microsoft's Clippy. And it has a whole package of security and privacy technologies to ensure your interactions are safe.

Most of all though, it's cute and about as sci-fi as it gets right now.

Of course, as it's just a proof of concept at present, there are no details on pricing, availability or specifications. Nor do we know what else it could be capable of, bar workflow help.

Still, it's good to see at least one company trying something new, and robots certainly do seem high on the agenda at the minute (just take a look at Honor's Robot Phone).

It'll certainly be interesting to see what else Lenovo (and others) can come up with for the next major tech trade show.

Google has shown off a new update to its AI image generation model Nano Banana 2, and it's looking fresh.

This is part of the Google Gemini AI sub-model that's specifically focused on the task of image creation based on requests.

The idea here is to bring the original Nano Banana model, released in August, more up to speed with the top-end Nano Banana Pro that landed in November. This latest Gemini 3.1 Flash Image, Nano Banana , is here to offer high-quality results at faster speeds.

What's new in Google Nano Banana 2?

One of the key offerings in this latest AI image generation model is speed. Since this now uses the high-speed smarts of Gemini Flash, image generation is faster than ever.

Crucially, this means quick iterations and edits can be done without lots of waiting. That should mean even if it isn't right first time, you can tweak and adjust to perfection quickly.

Accuracy is also important here as information is pulled from Gemini's real-world and real-time information base. That should mean more accurate renders with deeper understanding of what you're asking for.

Text is also more usable now, with precision addition, making this far more useful for marketing, greetings cards and mock-ups alike.

Creative control is improved with higher-quality photorealistic imagery available. Character resemblance can be maintained for up to 14 objects in a workflow. Better instruction understanding allows the model to follow commands, even when complex, with greater results.

You are also able to output to production ready quality, including up to 4K outputs, ideal for social posts or wide-screen backdrops. Thanks to more vibrant lighting and richer textures, details should be far more impressive also.

Google Nano Banana 2 is available to try and use now via the Gemini app, Google Search, AI Studio + API, Google Cloud, Flow and Google Ads.

Samsung has announced that it's going to offer direct access to another AI service, with the integration of Perplexity. That will allow Samsung users to say "Hey Plex" to trigger another voice assistant.

The justification for the move – announced via a press release ahead of this week's Unpacked – is to offer a "multi-agent ecosystem", because according to Samsung, "people are increasingly using multiple AI agents depending on the task".

This comes from Samsung own internal research, so there's no way to determine exactly who was asked and what the question was.

The inclusion of Perplexity will be an expansion of Galaxy AI, the system of AI features that Samsung is developing through its phones and other devices. Adding Perplexity means more choice, so if that's your AI agent of choice then you're going to well served.

Samsung goes on to say that the aim is to have Galaxy AI "curate experiences" from diverse services. As it stands, it's likely that Samsung device owners will have the choice of saying "Hey Google", "Hey Plex" and "Hey Bixby".

What will Perplexity offer on Samsung Galaxy S26 phones?

Perplexity will be deeply integrated into Samsung's apps such as Notes, Clock, Gallery, Reminder and the Calendar. This sounds a lot like the experience of Galaxy AI, where features like the writing assistant only work with Samsung's keyboard – and not with the popular GBoard from Google, for example.

While choice is great, there's a chance that this could lead to confusion. Samsung's not the first to offer multiple agents – that's something that's already in Moto AI – but the Motorola experience feels a bit disjointed, with Gemini usually offering a better universal option.

While the justification for the move from Samsung is given as creating a multi-agent experience, the move looks like the firm is ensuring that's it's not too dependent on one provider. We've seen this before, with Tizen developed as an alternative to Android software in the early days, the use of Exynos rather than Qualcomm, and so on.

Adding Perplexity removes the dependency on Gemini and gives Samsung the chance to create something that's unique – a brand experience that it can own. However, with Gemini accelerating at a great pace, there's every chance that users will choose Google over Galaxy AI.

The new Perplexity skills are expected to be available in One UI 8.5.

A new Google Gemini AI update has been announced which will enhance the artificial intelligence in a big jump forward.

Gemini 3.1 Pro has been announced by Google as the next iteration in the AI. Usually Google does a .5 upgrade and this will be the first with .1 at the end. Despite the smaller number jump, Google is clear this will be a big jump forward in terms of development.

Google says this new version is "for tasks where a simple answer isn’t enough," and that this upgraded core intelligence "represents a step forward in core reasoning."

In fact this, according to Google, "more than doubles the reasoning performance of 3 Pro." This is based on the ARC-AGI-2 score of 77.1% in testing.

So what does all this mean for us?

How will Gemini 3.1 Pro offer me more?

This new advanced reasoning system will be adept at offering practical use cases. Specifically this is designed to help in situations where a simple answer isn't enough.

So, rather than giving the most simple and concise answer, this version of Gemini can offer rich and complex outputs for you to work with.

Specifically Google says this works well when "you’re looking for a clear, visual explanation of a complex topic, a way to synthesize data into a single view, or bringing a creative project to life.”

So this can mean, quite literally, more rich image and video outputs when you want an idea visualized graphically.

When is Gemini 3.1 Pro available?

Google has begun rolling out the Gemini 3.1 Pro app to subscribers. It will also be available for Google AI Pro and Ultra subscribers as well as those using NotebookLM.

For developers, it's also available in the Gemini API via Google AI Studio, Antigravity, Vertex AI, Gemini Enterprise, Gemini CLI, and Android Studio.

Initially this will launch in preview mode, to help perfect everything, before rolling out on a wider scale in the near future.

Apple has been in the news a lot recently. We’ve seen several reports surrounding the next iPhone 18 Pro models and whether they will or will not ditch Dynamic Island in favour of under-display Face ID.

We’ve also heard that the regular iPad and iPad Air are about to get a performance boost, and then there’s been the constant chatter about more powerful MacBooks. That was even before the renders of the unannounced iPhone 17e appeared.

But the latest report from Bloomberg (via 9to5Mac) doesn’t mention any of that. Instead, seasoned Apple-expert Mark Gurman discusses three new AI wearables, including an AI pin/pendant, AirPods with cameras, and smart glasses that would compete with Meta’s Ray-Bans.

What could Apple's new AI devices offer?

According to Gurman, the products are being “built around the Siri digital assistant”, which makes sense following all the news about the smarter Siri arriving soon.

The AI glasses will apparently have speakers, microphones, and cameras, with one camera lens for capturing images and videos and the other for “computer vision” – to provide the glasses with “environmental context”.

"Early prototypes of the glasses connect via a cable to a standalone battery pack and an iPhone, but newer versions have the components embedded in the frame," he claims.

"The design uses high-end materials, including acrylic elements intended to give the glasses a premium feel. Apple is already discussing launching the device in additional styles over time.”

This supports a separate suggestion from The Information that Apple is working on a wearable AI pin, Bloomberg itself has also said this new device will be “reminiscent of the failed Humane AI pin”. Although it won’t have a display or projector, and is reported to rely heavily on the iPhone rather than being a standalone device.

It might also come with a necklace, to be worn in a pendant style.

According to Gurman: “The pendant would essentially serve as an always-on camera for the smartphone that also includes a microphone for Siri input. Some Apple employees call it the 'eyes and ears' of the phone.

"One area of debate for the product has been whether or not to include a speaker, which would allow users to hold back-and-forth conversations with the device directly. That means they could leave their iPhone in their pocket or bag or not wear AirPods.”

Not much was said about the AirPods featuring cameras, apart from that they could arrive this year.

For now, nothing is confirmed, though it certainly seems like there is plenty in the works, so watch this space.

The enhanced version of Siri that was first introduced by Apple in June 2024 may be facing continued internal delays but will still appear in 2026, according to Apple.

We first got a glimpse of Apple's plans for a more advanced Siri experience in 2024, when Apple Intelligence was announced. While Apple Intelligence has appeared on iPhone, iPad and Mac models, the enhancements to Siri remain ominously absent.

Through 2026, there has been quite a lot of reporting on the changes to come. First, there was the announcement of a partnership with Google, outlining that Gemini would power the next-gen Siri. It was expected that the first moves towards a smarter Siri would appear in iOS 26.4, but Mark Gurman reported on 11 February that there were further delays.

There's the suggestion therefore, that changes to Siri aren't going to appear in iOS 26.4, but in later versions. This, it seems, put the cat among the pigeons, with Apple stock dropping 5% on Thursday, according to CNBC.

However, CNBC also reported that Apple has confirmed that the updated Siri is "still on track to launch in 2026".

What we're left with is a slightly confusing picture. Initially, it was understood that new features would come to Siri under the Apple Intelligence banner with iOS 26.4 before a larger roll-out of the new Gemini-powered Siri experience with iOS 27 around September 2026.

With the report of delays, it's unclear which part of the changes to Siri are being referenced. But with the confirmation from Apple, we can still expect to see Siri undergo a radical change in 2026 to offer an enhanced experience, with better natural language understanding and capable of more chatbot-like interactions.

Why the delay to a more powerful Siri?

Siri was the first voice assistant, launching before Google Assistant and Alexa, but it has been slow to evolve, with the general feeling that Apple is playing catchup in the AI space.

Google has been a lot more aggressive with AI because it's investing heavily in the AI race – but that has also resulted in a fairly coherent AI strategy on Android phones.

With Siri in line to adopt Gemini's skills, Apple should be able to craft AI into meaningful experiences that some brands are currently missing. In many cases, we're seeing AI for the sake of AI, rather than because it's actually useful.

While Apple's relative caution in adopting AI has drawn plenty of criticism, there are high hopes that once it is fully engaged with delivering AI experiences, there will be more direction and purpose, rather than iterative versions of searching, summarising and translating.

I love the idea of being a watch designer. Putting your creative stamp on a small metal case, which houses a mechanical object so intricate that it draws comparison to the very world we live in, seems like an excellent way to join the annals of history.

In doing so, you'd become part of an esteemed group. Legendary watch designers like Gerald Genta and Jorg Hysek have an almost rockstar-like stature, which bridges generations of horological enthusiasts.

For me, there's just one problem – my skills with a pencil tend really don't go much beyond stickmen, so I'm not exactly best placed to go drawing up anything.

That's where the new Swatch AI-Dada tool comes in. This AI-powered platform allows users to input a prompt and generate a watch design right in front of their eyes, meaning your lack of artistic talent is no longer a barrier to your watch design dreams.

To find out if it's any good, I had a go at designing a watch, and got it shipped out to see the end product, too.

How to design a watch using the Swatch AI-Dada tool

First things first, you'll need to head over to the Swatch website and setup an account to gain access to the tool. That's all pretty quick, though, and then you're into the design side.

Thankfully, that's also really straightforward. Simply pop in a prompt of up to 300 characters and fire away. It takes about two minutes for the AI to cook something up based on your recipe, and you'll then be presented with a render of the design.

Once you like it, you can make a few more adjustments to things like the colour of the movements and the opacity of the hands, and then you're all done. If you don't like it, you can simply try again, with a maximum of three prompts per day, per account.

It's worth noting that my experience was ever so slightly different to that, as I was working on this before the tool had launched. Therefore, I had to send a prompt over via email, and didn't get to see the watch until it landed on my doorstep a few days later.

What is the watch like from the Swatch AI-Dada tool?

First things first, let's talk packaging. It's an often overlooked part of the experience, but you can really tell that every effort has been made to create an event of this moment.

Pulling off the outer AI-Dada branded sleeve reveals a Swatch branded box inside. That opens like a book, revealing your creation inside. It feels really momentous.

The watch itself will feel instantly familiar to anyone who has used a classic Swatch model. The strap is very supple, and the plastic case ensures complete lightness which is great for going unnoticed as you wear it.

One thing I will note is that it's an incredibly loud watch – in fact, I could hear its quartz movement ticking before I'd even cracked into the box! If you're someone who keeps a watch on the nightstand, that's worth bearing in mind, though mine are all stored away from where I sleep, so it's not an issue.

There's also the issue of cost. At £163 (approx. €187 / US$222 / AU$312), this is just a little more than double the price of the equivalent model in the standard range. That's going to be a tough pill to swallow for some, with the model clearly bearing the cost of its creation tool.

Still, I do think it represents good value – particularly if you're trying to design something very specific. Ultimately, that price is still relatively cheap in the world of watches, and for a little piece of fun, I'd say it's worth it.

Alexa Plus is now available to everyone in the US, following an early access period that's been running since the new AI-powered assistant was announced in February 2025.

Alexa Plus is an evolution of the familiar voice assistant that's been in Amazon's Echo devices since they launched in 2014. Since that time, it has evolved into one of the dominant voice assistants, especially in the home, where it's also common in TVs and other devices.

The move to Alexa Plus is designed to give the service greater powers, shifting to natural language understanding with the aim of being more useful. Amazon envisions users having long conversations with the AI, mining for information, making decisions and ultimately, using Alexa Plus to get things done.

That's going to include the likes of arranging a party, pulling together the menu and inviting participants, with Alexa assuming an agentic role – ie., where the AI can take actions on your behalf.

But it's taken a while to be fully available.

Thankfully, Amazon has now confirmed that Alexa Plus is open to everyone in the US, meaning it's stepping out of that limited access and moving into the mainstream. It's free for Prime members, which will help adoption, while early reviews have been generally positive.

Amazon launched the Echo Dot Max, new Echo Studio and recent Echo Show models specifically with Alexa Plus in mind, with more power available to make the AI assistant more seamless. However, the service is accessible through exisiting Echo devices too.

The Echo Show comes with the advantage of providing a visual element to proceedings, which is likely to make Alexa Plus more useful and easier to engage with. It will also be available through a browser and in the Alexa app.

While Alexa has enabled many smart homes over the past decade thanks to wide support for third-party devices, Alexa Plus promises to enhance the experience, allowing users to say what they want to happen, rather than having to stick to rigid language – including setting up Routines (home automation) on the fly.

The move to Alexa Plus reflects the changing nature of such assistants, with Google pushing Gemini across a full range of its devices, while it's now been confirmed that Siri on Apple devices will be enhanced using Google's Gemini tech too.

Alexa Plus is also still absent in much of the world, having only launched in the US and Canada so far. It has been promised for a launch in the UK in 2026, but things are taking a little longer than perhaps expected.

The challenges to Alexa Plus don't stop there, with the likes of ChatGPT having a stranglehold on smartphones while Gemini is also widely accessible. The question is whether Alexa Plus can make itself impressive enough on Echo devices for users to choose to use it on a phone and keep all their AI interactions in one place.

For now, if you're in the US and a Prime subscriber, then you should have access to Alexa Plus. For everyone else, the wait goes on.

Gemini might soon be able to help with a range of tasks thanks to a system called "screen automation". Code referring to these action has been found in the Google app, leading to more understanding of the company's plans for the AI system on Android.

Google announced Gemini as the replacement for Google Assistant in 2023, with the AI-powered assistant making big strides in 2025 to bring smarter interactions to your devices. But for many, Gemini has really found its place on the smartphone, while others wait for its arrival on TV and in the car.

Google has always talked about the agentic powers that it wants AI to offer – that is the ability to do things on your behalf. It sees a future where you'll ask Gemini to create an itinerary for a weekend away and once you've approved the suggestions, you'll get the AI to book hotels and make reservations at restaurants.

It's these latter actions that are seen as "agentic", ie., the AI is acting on your behalf, beyond just surfacing information for you.

According to 9to5Google, there are strings of code in the Google app 17.4 beta that refer to "Get tasks done with Gemini". It appears to be codenamed "bonobo", with Google explaining that it can help with "placing orders or booking rides, using screen automation on certain apps".

Using screen automation suggests that this isn't going to happen in the background, it's likely that it's all going to happen in front of your eyes, meaning you can intervene if things aren't heading in the right way.

How could an agentic Gemini help you?

While there's no working example, it's easy to imagine that asking Gemini to book you an Uber home would open the Uber app, fire up the destination, select a ride and tell you when it's arriving.

Equally, asking Gemini to order your usual from a delivery service should be fairly straightforward, as long as Gemini knows your preferences.

The source identifies a couple of additional pieces of information lurking in the app, including the declaration that "Gemini can make mistakes" and that "you're responsible for what it does on your behalf".

This latter phrase is a great disclaimer, so that if Gemini orders a stack of pizzas, you can't then turn around and say that you didn't want them and the AI is to blame. I think we all know that's going to happen a lot in this agentic future.

It's said that you'll be able to intervene and stop the AI at any point, but that does sort of beg the question of why you'd automate this process in the first place? Is it so arduous to open an app (which Gemini will already do for you) and make a few taps?

It's said that "screenshots are reviewed by trained reviewers and used to improve Google services if Keep Activity is on" which will raise privacy questions with some users, while it also says "don’t enter login or payment information into Gemini chats", suggesting that it can't be trusted with sensitive information.

Like all things AI, the big question comes with whether this is a good thing or not. We know that AI is gaining skills all the time, with the ultimate aim to carrying out tasks on your behalf in the future.

But if it can't be trusted to get it right and there's a chance that personal information might not be secure, is this ever going to get off the ground?

If you've ever dreamed of designing your own, one-of-one, unique watch, today is your lucky day. That's because the good folks at Swatch have unveiled a new tool, designed to do just that.

The AI-Dada tool isn't exactly new, having first launched back in November last year. Back then, it was only available to users in Switzerland, though, meaning many of us simply had to watch from the sidelines.

As of today, the service is available in a further 20 nations around Europe, including here in the United Kingdom. The full list of added countries is as follows:

• Austria
• Belgium
• Denmark
• Finland
• France
• Germany
• Greece
• Ireland
• Italy
• Liechtenstein
• Luxembourg
• Netherlands
• Norway
• Poland
• Portugal
• Slovakia
• Spain
• Sweden
• Turkey
• United Kingdom

How to use the Swatch AI-Dada tool

To find out if the tool is any good, I had a go at designing some watches on it. The process to get set up is incredibly simple – you'll need to create an account with a name, email address and password, but then you're good to go.

The tool itself is incredibly easy to use. Once in, you'll see a prompt box, where you can enter your wishes for the design. That allows for up to 300 characters, though I found myself naturally only using about half of that.

Still, you'll want to be as specific as you can, as the tool will only allow users to make three designs per day. That's presumably in a bid to combat excessive use of the AI platform, and keep it accessible for as many people as possible.

I used a different prompt for each of my three designs, and I was very impressed with each of the results. The overall designs in each instance were much closer to what I had imagined when entering the prompt than I thought would be possible.

We'll have a fuller review of the process with an AI-designed watch to review in the coming days.

After announcing it at IFA 2025, SwitchBot has finally launched its new AI Art Frame. Dubbed as the ‘world’s first’ AI-powered Color E-Ink Art Frame, the SwitchBot AI Art Frame displays historical, renowned artworks, and allows you to make your own – but the AI Art Remix feature is the one you’ll love the most.

Available in 7.3-inch, 13.3-inch and 31.5-inch sizes, the SwitchBot AI Art Frame has a similar concept to smart displays when they go into standby. Rather than looking at a plain black screen, some smart displays have an option to display artwork so you have something nice to look at while it’s not in action.

The SwitchBot AI Art Frame puts art at the forefront of your home, but rather than just looking at one picture, you can switch it to display various works of art from the likes of Vincent Van Gogh, Leonardo Da Vinci, and other popular artists. It sits in an aluminum frame, can show pictures horizontally and vertically, and is wireless so you can mount it on a wall or prop it on tables.

With E-Ink Spectra 6 technology, the SwitchBot AI Art Frame mimics the texture of the art it displays so the paintings look true-to-life. The display is also glare and backlight-free so you can see all the colours and details in clear resolution. As a wireless frame, the built-in battery offers up to two years battery life per charge and only uses power when it changes pictures.

So, what’s the AI part? Powered by NanoBanana, the SwitchBot AI Art Frame has AI features including AI Art Prompt and AI Art Remix for users to create their own artwork. With AI Art Prompt, users can enter textual prompts or visual assets, and upload sketches which the SwitchBot AI Art Frame uses to create unique pieces of art.

But it’s the AI Art Remix I’m most excited about. AI Art Remix allows you to make your photos or sketches into pieces of art, and you can choose a specific style like oil painting, water colour and anime. As someone who’s absolutely terrible at art, AI Art Remix is definitely something I can get my head around and would love to play with.

The SwitchBot AI Art Frame is compatible with the SwitchBot app which is where you’ll want to upload your content. Users can also schedule different art for events, holidays and time of day, and you can store up to 10 pictures locally.

I got to see the SwitchBot AI Art Frame at IFA 2025, and I thought it was a really cool concept. However, at the time, I felt it needed a bit more development as it was a bit glitchy when changing photos. It seems that this has been worked on though, and the addition of AI features makes it a much more exciting prospect, and one I can see budding artists using.

Available to buy now, prices on the SwitchBot AI Art Frame starts at £149.99 / $149.99. As of writing, shoppers can already get it with 20% off for a limited time. The SwitchBot AI Art Frame also comes with a subscription plan for utilising certain features which is priced at £3.99 / $3.99 a month.

We’ve covered Gemini rather a lot this month. From the news that it will be powering the smarter Siri due to arrive in a matter of months, to the new auto-browse feature in Chrome that'll see it act as a more personalised assistant, it has dominated headlines.

Now we're about to see something that takes its capabilities to a whole new level.

After revealing project Genie 3 in August 2025 as a world model "capable of generating diverse, interactive environments”, Google has started to roll it out for AI users.

What is Project Genie?

On its The Keyword blog (via 9to5Google), Google has detailed what Project Genie is capable of through a series of videos. In a nutshell, if you can imagine something, you can both create it and play within it.

The software will require you to simply add a prompt to the "Environment" box describing what you want your world to look like. You’ll then describe what you want your character to be in the "Character" box next to it.

Once done, tap on "Create Sketch" and that's it. There’s no need for any coding, you’ll just need your imagination and a good description or image for each step.

As an example, Google uses this description for the environment in the video showcasing Project Genie in its blog post: “A photorealistic alpine meadow with wildflowers. Among the evergreen pine trees is a rustic log cabin with a front porch...”

For the character, it used: “A shiba inu centered in the frame, angled like a 3rd person video game, with highly responsive controls.”

After the sketch is created, another screen appears where you can then tap on "Create World" to allow you to play it using the W, A, S, D and Space keys.

Always fancied climbing Everest? Create Everest as the environment, you as the climber and away you go with no need to leave your warm and cosy sofa.

There are just a couple of catches. Firstly, you will beed to be a Google AI Ultra subscriber in the US and secondly, you’ll need to be okay with these known areas for improvement as described by Google in the blog post:

“Generated worlds might not look completely true-to-life or always adhere closely to prompts or images, or real-world physics. Characters can sometimes be less controllable, or experience higher latency in control. Limitations in generations to 60 seconds.”

You’ll also need to be over the age of 18.

It was only the other week that we covered Google finally bringing Gemini to Chrome in ChromeOS, but there's some much bigger news today.

In what is arguably the biggest shake-up to the Chrome browser since its launch in 2008, Google is introducing a new feature – auto browse. It is designed to transform the browser into a fully-fledged personal assistant to help you get things done.

Powered by Google’s Gemini 3 model, auto browse isn’t about faster searches, but instead, about putting Chrome in the driving seat to “handle multi-step chores on your behalf”.

What is Chrome auto browse and when can you get it?

Instead of you spending twenty minutes jumping between tabs on Chrome to compare flight prices or fill out a form, auto-browse will be able to complete these tasks for you.

As one example, the company said the feature is able to help you plan a holiday by researching hotel and flight costs across multiple date options for you. It will then suggest a budget friendly option that suits you. It will also be able to do more complex tasks too.

In its blog post, Google said: “Let’s say you’re planning a Y2K theme party and find inspiration in a photo that perfectly captures that era. Auto browse, with the multimodal capabilities of Gemini 3, can identify what’s in the picture, search for similar items and add them to your cart – staying within budget and even applying discount codes.

“And if you give auto browse permission, it can use Google Password Manager to handle tasks even if a sign-in is required.”

The company also revealed that its testers have used the feature for “scheduling appointments, filling out tedious online forms, collecting their tax documents, getting quotes for plumbers and electricians, checking if their bills are paid, filing expense reports, managing their subscriptions and speeding up renewing their driving licenses". Phew.

Essentially, it could be key for completing some of those life admin tasks you put off.

In addition to the auto browse feature, Google has introduced a side panel for Gemini in Chrome, allowing for easier multi-tasking. Connected Apps are also being supported with the update.

It means Gemini will integrate with Gmail, Calendar, YouTube, Maps, Google Shopping and Google Flights to help you get things done quicker.

The new features are rolling out now in preview for Google AI Pro and Ultra subscribers in the US, across Windows, macOS, and Chromebook Plus.

Sadly, there's no word yet on when they might arrive in the UK, but we're hoping it'll be soon.

Back in November, Google Maps got an upgrade we'd been waiting a long time for. That saw the powerful Gemini AI integrated into the platform, with a view to making on-the-go assistance even more seamless.

That was great for drivers, meaning they could engage in a natural conversation to gain information and keep their attention focused on the road. Now, that functionality is rolling out to even more users, with functionality to help those who are walking and cycling.

The idea for cyclists appears to be much the same as for drivers – offer a solution which is as seamless as possible, ensuring you can keep your focus on the road, rather than on your phone. You'll be able to ask it all of the usual questions, like "What's my ETA?", "When is my next meeting?", or "What is the meaning of life?", all while keeping two hands on your handlebars.

For me, though, the implementation of the feature while walking could be really handy. I travel a lot, and often find myself getting lost in the middle of a city I've never been to before – it's all part of the fun, right?

With this new feature, I would be able to ask Gemini where I was, what there is to do in the local area, a good spot to get lunch and even find a good spot for a photo opportunity nearby. The possibilities are endless, really, and that could make it an invaluable tool for frequent flyers, or just those who get lost easily.

There's good news for iPhone users, too. The feature is live now around the world for those using Google Maps on iOS – as long as you're in a location where Gemini is supported, of course.

Those on an Android phones may have a little more of a wait, as the feature is rolling out for those handsets as we speak.

When you think of image manipulation, there is a very good chance Adobe’s Photoshop springs to mind. Even if you don’t use it yourself, you only have to have seen Crazy Stupid Love to understand the software’s influence when it comes to creating "perfect" images.

“Seriously? It’s like you’re Photoshopped,” is the classic line said by Emma Stone to Ryan Gosling when he removes his shirt – and to be fair, he does.

But that film came out in 2011 and 15 years later Adobe thankfully hasn’t rested on its laurels. The company has just dropped an update that promises to make editing workflows smoother, sharper, and, a little bit like magic (as if it wasn’t already).

If you thought the Generative Fill feature was impressive before, this latest update takes things to a whole new level.

What’s new for Photoshop in 2026?

There’s been a huge boost to the Firefly-powered tools, for starters.

Generative Fill and Generative Expand now support 2K resolution output, meaning AI-generated additions will look crisper and blend more seamlessly.

Adobe has also refined the lighting and depth algorithms, so those awkward seams and AI-looking artefacts will start to become a thing of the past.

There is also an upgraded Reference Image feature. It’s now what Adobe is calling “geometry-aware", so if you drop a reference object into a scene, Photoshop automatically calculates the correct scale, rotation, colour and perspective.

The update introduces a few new adjustment layers, too.

Clarity, Dehaze, and Grain can now be applied as maskable layers, allowing for precise, reversible edits that were previously much more fiddly to manage.

Last but certainly not least (I think this one might be my favourite), the new Dynamic Text (in beta) allows you to warp text inputs into circular or arched shapes with a single click, automatically resizing to fit paths without the usual workarounds.

The Photoshop update is rolling out now to all Creative Cloud subscribers. You should see the new features appear imminently.

Who doesn’t love a meme? They are perfect for a chuckle, whether you come across them on whatever social media platform you might be doomscrolling at the time, or a friend sends you one knowing it will make you laugh. More often than not, they're guaranteed to brighten someone’s day.

Well, now you don’t have to wait for the perfect meme to appear – you can turn yourself into one. Google has announced a new feature on a Google Photos support page (via 9to5Google) called Me Meme.

It will allow you to use generative AI to “help you create funny, personalised images” – according to the Google support page.

“It's a simple way to explore with your photos and create content that's ready to share with friends and family,” it adds.

How to create a Me Meme with Google Photos

So, how do you create one? Well that part is actually pretty simple, as long as you live in the right place.

It’s worth noting before we go too much further that the Me Meme feature is currently only available to those in the US, though hopefully it will roll out to other regions soon.

If you are in the US, you will need two things. First, you'll need a template.

You will be able to choose different options from those available, or you can upload your own funny picture to use.

And second, unsurprisingly, is a photo of you – you just need to select the image you want to use from Google Photos.

To create your meme, you then need to follow these steps:

1. On your Android device, open the Google Photos app
2. At the bottom, tap Create
3. Then tap Me Meme. Note, if it’s your first time, you can follow the on-screen instructions after the intro screen
4. Select a template
5. Tap Add photo
6. Select the photo you want to upload. Google suggests “for potentially better results, try to upload a well-lit, focused, and front-facing portrait photo”
7. Tap Generate

After the meme is generated, there are a couple of extra steps to follow:

1. To save the meme: Tap Save
2. To generate again: Tap Regenerate
3. To share on the other platforms: Tap Share

That’s it! Maybe you’ll even go viral and bring a smile to multiple people.

Apple's plans to make Siri more like a chatbot have been revealed, with the company looking to fend-off the popularity of AI services like ChatGPT. The new version of Siri will reportedly be powered by Gemini, which already offers a chatbot like experience on Android devices.

The information about Apple's plans comes from Mark Gurman at Bloomberg, an established source for details about forthcoming Apple plans. The chatbot, which has the codename Campos, will replace the Siri interface, making for a conversational AI experience supporting both typing and voice.

Details about the new version of Siri come following the announcement that Apple and Google are entering into an AI partnership, with the details of the new Siri sounding just like the services currently offered by Gemini.

Apple made its first foray into AI with the announcement of Apple Intelligence in 2024. However, there's been a muted response to the efforts by the Cupertino tech giant, with Siri remaining well behind the capabilities of services like Gemini.

While Apple had outlined that Apple Intelligence would allow users to draw in third-party services like ChatGPT, the more sophisticated version of Siri that was promised has remained ominously absent.

Rivals, meanwhile, have forged forwards, with Gemini particularly gaining a wide range of skills and broadening its offering. Taking that technology into a Siri package, deeply integrated into Apple devices makes perfect sense and would bring Apple back into line with its competitors.

According to the source, Apple will retain the freedom to switch to other AI models in the future, so it's not wholly dependent on Google.

It's reported that the new Gemini-powered Siri will be able to create content and images, provide summaries, search the web for information and analyse uploaded files. It will also be integrated into more personal data so it can help locate files, messages, images, events and more.

It will also be integrated into core Apple apps, bringing the sort of advantages that Gemini offers across Google's apps. However, it's thought that Apple will want to limit the memory of the assistant, to protect user privacy.

While Siri is expected to be updated in iOS 26.4 in the coming months, the new Gemini-powered Siri is reportedly going to launch with iOS 27, in line with new iPhone models in September 2026. It's expected to be previewed at WWDC in June 2026.

It feels like I was in Vegas just yesterday, but we're already approaching a week since CES 2026 was actually in full flow, and I'm catching up on the very final parts of what I was able to see in a jam-packed few days of tech launches and showcases. Sadly for Razer, I haven't saved the best for last – but I can certainly say its suite of new products and concepts had memorable parts.

Razer made its name (and continues to do so) as a gaming accessory brand, with gaming mice and headsets really acting as its heartland, but it's branched out quite heavily in recent years. After successfully establishing itself as the maker of great, sleek laptops, it now seems to have been sipping from the AI cup, leading it to a couple of frankly baffling concept showcases at this CES.

The first, and the one that will actually be available to buy soon (apparently, since you can pay a refundable $20 to reserve one now) is a desktop upgrade for its AI companion software, Ava. This little tube will sit on your counter, presumably next to your PC, and project a small holographic figure internally, with a design chosen from a roster of extremely gamerish options.

The one I saw was predictably pretty anime-inflected in its stylings, although slightly more buttoned-down options will be available, too. Still, even after watching a demo of how the AI could react to on-screen action with tips and even criticisms, I just can't understand Razer thinking that it'll have widespread appeal.

Sure, there's an angle where some coaching while you game could help you improve, but from what I saw that's quite a long way from being what Ava can offer. The projection is the really new bit here, meanwhile, and it's frankly just a bit too gamerish to win me over.

Then there's the elephant in the room – Razer confirmed at the show that Ava was running on a version of Grok, xAI's chatbot. You couldn't pick a more controversial time for that information to come out, given the rightly radioactive controversy swirling around Grok's nudification features, which stand to see it potentially banned in various countries.

It's a great summary of why AI launches can be so perilous and ill-conceived – two words that sadly also describe Razer's other big swing this CES, Project Matoko. I was told very earnestly about Razer's belief that smart glasses are a bad idea, with too many compromises, and that a better answer to bringing AI assistants with you is to house it in some headphones.

Hence this more distant concept, Matoko, which sees a Razer gaming headset modified with the addition of two front-facing cameras. You can talk to, and ask questions of an AI companion that you activate when wanted, presumably using a very similar model to Ava (although this wasn't confirmed).

The hardware was perfectly decent here, and lightweight enough to be comfortable. I can even see how Razer can argue it's a more seamless integration into a headset than the equivalent in glasses (which get chunkier and heavier to accommodate the extra tech). All-day battery life is a plus, for example.

However, I just can't see how this product becomes a reality that people really want. Perhaps (and it's a big perhaps) if a brand like Apple launched a new version of AirPods Max with cameras, then people might get on board. But Razer? The gaming peripheral brand, becoming a provider of AI-enabled all-day headphones for regular use?

I just can't see how that makes sense at all, from a brand or a market perspective, and it makes it really tough to believe that this will ever really be widely available – or sell more than a relative handful of units.

In some ways, these concepts are a microcosm of the AI boom, in that Razer is far from the only tech brand trying to go big on AI and failing to win over actual customers. I'm very prepared to eat my hat if Project Matoko goes on to open up a new category of products, but for now I feel pretty confident that Razer will be better off sticking to what it knows better.

Google is certainly in the lead when it comes to consumer-facing AI. Not only has it recently announced a ground-breaking deal with Apple for Gemini to prop up the next generation of Siri and Apple Intelligence in general, it has unveiled an upgrade coming to its own services that could change the way we use them forever.

Personal Intelligence is an overarching system that can link the Google apps and services you use daily, to provide better assistance in each. It'll initially be available to US users in the Gemini app and in beta form, and will work with Gmail, Photos, Search and YouTube for more natural responses to your requests.

This includes proactive problem solving. One example Google gives is if you're searching for new car tyres, the Personal Intelligence engine could suggest specific types based on your actual car model, taken from insurance documents it finds in Gmail.

Alternatively, you might get more tailored travel recommendations based on recent bookings you've made and interests gleaned from your Photos library.

Now, this might all seem intrusive – and it's hard to argue otherwise. But Google is keen to stress that this is an opt-in service. It is switched off by default. It will also be initially restricted to subscribers of Google AI Pro and AI Ultra.

It's not restricted to Android though – once activated, it'll work across Android devices, but also iOS and the web.

Users of the free tier and in other regions will be able to access it "soon".

"The best assistants don't just know the world; they know you and help you navigate it," wrote Google vice president Josh Woodward for the company's official blog.

"Today, we’re answering a top user request: you can now personalise Gemini by connecting Google apps with a single tap."

How to turn on Gemini Personal Intelligence

To opt into Personal Intelligence, you need to subscribe to Google AI Pro or AI Ultra and be in the US.

If so, just head to the latest version of the Gemini app, tap Settings, tap Personal Intelligence and select the apps you are happy for it to use (ie. Gmail and/or Photos).

The service will then consider content within those apps when personalising responses to your Gemini requests.

If you're in the market for a new phone, you might be surprised by the cost of some models out there. Prices have been steadily rising over the last few years, making owning high-end devices less of a possibility for most people.

Now, Carl Pei – former co-founder of OnePlus and current CEO of Nothing – has taken to Twitter to explain why. In the article, Pei explains the memory shortage which has plagued the tech industry for the past few years.

That short supply means that prices have also gone up. Pei says, "Modules which cost less than $20 a year ago could exceed $100 by year-end," which showcases some of the difficulties manufacturers are facing.

He goes on to suggest that the landscape for brands is stark. Manufacturers must either "raise prices, by 30% or more in some cases, or downgrade specs," he suggests.

A new era for phones?

Pei goes on to talk about how Nothing has been built for this environment – one where the race for more specs at a lower price gives way to a design-led preference which focuses on the user experience. And honestly, I don't think he's too far off the mark.

Users will always look for some spec improvement when upgrading their handsets, but the pace at which those improvements have come to market feels like it has slowed of late. It's not uncommon for users to hang on to their handsets for three to four years now, where two was very much the norm a few years ago.

There's no telling quite how this will all play out. Personally, I hope that the need for something different pushes the envelope, progressing the technology we have available to us.

In the short term, though, I suspect more and more of us will be turning towards the more affordable alternatives from large brands and interesting products from challenger brands in a bid to restore a sense of value in what we use.