Feed aggregator

• The Samsung Galaxy Z Fold 7 could be just 3.9mm thick when unfolded
• It might also have new battery and charging tech
• We're also hearing again that the Galaxy Z Flip 7 could have a bigger battery than its predecessor

We’ve heard from several sources that the Samsung Galaxy Z Fold 7 could be impressively slim, but a new claim suggests it will be even thinner than we’d previously heard, at least when unfolded.

According to @UniverseIce, a leaker with a solid track record, the Galaxy Z Fold 7 will be just 3.9mm thick when unfolded – that’s even slimmer than a 4.5mm claim we’d previously heard.

When folded, it’s a slightly different story, with this source claiming the Galaxy Z Fold 7 will be 8.9mm thick, which would actually be thicker than the 8.2mm we’d heard elsewhere. Either way though, that’s exceptionally slim for a foldable phone.

According to some exclusive information, the next direction of Samsung flagship machine is still to be thin and light, and the battery will be thin and thin, and the body will continue to be thin. The Galaxy Z Fold7 will be the thinnest folding machine in the world at that time,… pic.twitter.com/uAs3NzAf8jMay 7, 2025

@UniverseIce also shed some light on the battery, echoing previous leaks in saying that the Galaxy Z Fold 7 will have a 4,400mAh battery, just like the Samsung Galaxy Z Fold 6, but intriguingly also claimed that it will use new battery and charging tech.

It’s unclear what they mean by this, but it could mean the battery will last longer than the battery in the Z Fold 6, despite apparently being the same capacity, and it may also mean we see higher charging speeds than the 25W of the Z Fold 6.

From the Fold to the Flip, and the S26

Moving away from the Galaxy Z Fold 7, the same source has also claimed that the Samsung Galaxy Z Flip 7 will have a 4,300mAh battery. That would make it bigger than the 4,000mAh battery in the Samsung Galaxy Z Flip 6, and this is something we’ve heard from multiple sources, so there’s a good chance it’s accurate.

Finally, @UniverseIce claims the Samsung Galaxy S26 “will continue to be thin”. For reference, the Samsung Galaxy S25 is 7.2mm thick, so this suggests its successor won’t be much thicker, and could possibly be thinner.

Of course we’d take all of this with a pinch of salt, but as far as the Samsung Galaxy Z Fold 7 and Samsung Galaxy Z Flip 7 go we should know all about them soon, as they’re likely to be unveiled in July. The Samsung Galaxy S26, though, probably won’t land until early 2026.

You might also like

• Should you wait for the Samsung Galaxy Z Fold 7?
• Five things we expect from Samsung Galaxy Unpacked
• Samsung Galaxy Z Fold 7 colors: every rumored and predicted shade

The Federal Reserve will likely hold interest rates steady Wednesday as President Trump's tariffs threaten to raise prices and drag down economic growth.

(Image credit: Kevin Dietsch)

The origin story of Alcatraz's name is a complicated mix of language, a history that dates back to the first Europeans to explore the coast of California, and, a flock of seabirds.

(Image credit: Justin Sullivan)

Senior presidential adviser Kari Lake says Voice of America will rely on coverage from the far-right OAN network. OAN has offered unwavering support of President Trump.

(Image credit: Jose Luis Magana/AP)

What happens when people put their phones down and eat together?

(Image credit: Marsden Shetterly)

Diminishing access to four-year universities and majors has been holding rural students back. Community colleges say they can bridge the gap.

(Image credit: Mike Rundle for The Hechinger Report)

• Microsoft has revealed some smart new features for Copilot+ PCs
• Windows 11 is getting an AI agent to help you easily adjust settings
• There's a bunch of other features coming (in testing), including powering up Narrator, the Photos app, and Snipping Tool

Copilot+ PC owners are getting some very useful introductions for Windows 11, including functionality that Microsoft has been promising for a long time now – the ability for AI to change settings for you in the operating system.

Microsoft covers all the introductions it’s making in a blog post – note that these additions will all be in testing only (for now) – as well as recapping a bunch of other features that have already been seen in testing.

The key introduction for many (coming off the back of the full release of Recall, finally – and more besides) will be the change for the Settings app which brings in an AI agent.

This takes the form of a bar at the top of the Settings panel into which you can type natural language queries. The idea is to take all the pain out of tweaking the various settings in Windows 11, of which there are a ton (and a half).

So, as an example, if you wanted to make the mouse cursor larger, rather than searching for ‘mouse’ in the normal Settings search bar, you can ask the AI instead, and just type a query: “How do I make my mouse pointer bigger?”

You can phrase it any way you want (within reason), and the AI should surface the option you need so you can click to enable it. Or it might go even further and make suggestions as to what setting you might want to pick (like a recommended size for the pointer in the example above). When it comes to such recommendations, the obvious caveat that AI can be wrong is included.

The ability for AI to make sweeping changes in Windows 11 has been something Microsoft has been talking about since it first introduced Copilot to the OS. And while these are smaller tweaks, rather than big swathes of options being adjusted, it’s still a very nifty move – and likely a safer way to work (for now, certainly) given the propensity for AI to make mistakes at times.

Check out the YouTube video above to see the feature in action. However, do note that this functionality is coming to Copilot+ PCs with Snapdragon X chips only to begin with (in testing, as mentioned), but it shouldn’t be long before it arrives for devices with AMD or Intel processors.

There are a few other notable features being brought in for Copilot+ PCs here (also in testing), and I’ll take a look at them next.

1. Relight your photos

The Photos app in Windows 11 is getting dynamic lighting controls which are facilitated by AI (and the NPU on your Copilot+ PC).

This allows you to correct poor lighting in an image, or just play around with different (or even oddball) lighting effects, using up to three light sources.

Check out the example tweaking shown in the above video (where another feature, object select in Paint, is showcased too).

2. Easy screenshots with AI doing the heavy lifting

The Snipping Tool is getting a new ‘Perfect Screenshot’ feature. Invoke this and you can roughly frame an area of the screen that you want captured, and AI will crop it down to the key part that (it assumes) you want to screenshot.

Check it out in the above sample video clip – this could be another substantial timesaver. Should the AI get it wrong, there are handles to adjust the errant crop, too.

3. Narrator gets powered up

(Image credit: Microsoft)

Windows 11’s screen reading tool is getting a boost, as Microsoft explains: “Narrator now offers rich image descriptions, which provide detailed context for visuals including charts, photos and UI elements.”

What this means is that where web content doesn’t have alternative text provided – words that can be read out to describe an image for those who can’t see it – Narrator can do this itself.

This should be a major step forward with accessibility for blind (or low-vision) Windows 11 users. As with the Settings AI agent, this functionality is coming (in testing) to Copilot+ PCs with Snapdragon X chips before other models.

You might also like...

• Been hiding from Windows 11 24H2 due to the fuss about all the bugs? There’s nowhere to run now as Microsoft’s made the update compulsory
• Microsoft could be working on the next update for Windows 11 – but 25H2 could end up being a disappointment
• Windows 11 fully streamlined in just two clicks? Talon utility promises to rip all the bloatware out of Microsoft’s OS in a hassle-free way

The superpowers have been locked in a geopolitical blinking contest, waiting for the other side to reach out. The talks in Switzerland are the first concrete sign of a potential thaw in the deadlock.

(Image credit: Ng Han Guan)

Bang & Olufsen Beosound A1 3rd Generation: two-minute review

The latest version of the Beosound A1 by Bang & Olufsen is the best yet – and when you consider how well-regarded the previous two versions were, it follows that this 3rd Gen is something special.

As you might expect from this brand (and as you’re likely to demand at this price), the A1 3rd Gen looks and feels like a high-end product. There's a heated debate among TechRadar's US and UK staff over whether the A1 Gen 3 looks more like a muffin, a crumpet or a bombolone, but whichever baked (or fried) breakfast (or teatime) treat works best for you also works for us. Yes, it looks like that, but in the most premium way. The materials used in its construction are relatively lavish, and build quality is unarguable. Its modular design means it can be repaired, and its battery can be replaced – both of which are good things and certainly help it all the way into our roundup of the the best Bluetooth speakers at its size and weight.

And between multipoint connectivity, aptX Adaptive compatibility, 60 watts of power, an IP67 rating (meaning it's up there with the best waterproof speakers too), plus a USB-C socket that can be used for data transfer as well as for charging, the A1 3rd Gen is pretty well-specified too. Only a lack of Auracast compatibility sounds a mildly false note.

Best of all, though, is the sound. The Beosound is a robust but deft listen, detailed and dynamic in equal measure, and able to present recordings with a nice open character at the same time as ensuring they sound unified and together. It’s short of ultimate scale, of course – a product of this sort of size is almost bound to be – but in every other respect it’s an engaging, revealing and thoroughly entertaining listen.

(Image credit: Future) Bang & Olufsen Beosound A1 3rd Generation review: price and release date

• $349 / £299 / AU$639
• Launched on May 6, 2025

The Bang & Olufsen Beosound A1 3rd Gen is on sale from today (7th May 2025) and in the United Kingdom it’s priced at £299. It’s $249 in the United States, and in Australia it goes for AU$639. Which means that no matter the territory in which you’re shopping, it’s priced like a premium device – which, in turn, means it’s Bang & Olufsen business as usual.

(Image credit: Future) Bang & Olufsen Beosound A1 3rd Generation review: specs Bang & Olufsen Beosound A1 3rd Generation review: features

• Bluetooth 5.1 with aptX Adaptive compatibility
• 60 watts of Class D power
• USB-C can be used for data transfer

The Beosound A1 3rd Gen uses Bluetooth 5.1 for wireless connectivity, and it’s compatible with SBC, AAC and aptX Adaptive codecs. That Bluetooth standard means it won’t support Auracast, though multipoint connectivity means your speaker can be connected to two sources at the same time. Once your digital audio information is on board, it’s dealt with by a digital-to-analogue converter of unspecified type and resolution, and then served up by a two-driver array. A 15mm tweeter and 83mm mid/bass driver each receive 30 watts of Class D power, in an arrangement Bang & Olufsen claims is good for a frequency response of 54Hz - 20kHz.

The adoption of aptX Adaptive is a step forward for the A1, but this 3rd Gen model dispenses with Amazon Alexa voice control and wider smart-control integration you'll find in the 2020-issue Beosound A1 Gen 2. Whether or not you consider this to be a backward step depends almost entirely on how much you enjoy the sound of your own voice.

Battery life is a very creditable 24 hours from a single charge (provided you’re operating at reasonably modest volume), while from ‘flat’ to ‘full’ takes a rather leisurely three hours. Charging happens via the USB-C slot on the edge of the speaker, and for this 3rd generation this input can also be used for data transfer – so the Beosound can be hard-wired to a source of music. It can also form a stereo pair with another A1 3rd Gen or, happily, an A1 2nd Gen should you have one.

• Features score: 4.5/5

(Image credit: Future) Bang & Olufsen Beosound A1 3rd Generation review: sound quality

• Impressive low-frequency extension and substance
• Consistent and balanced even at volume
• Great detail retrieval and midrange resolution

With one or two fairly minor caveats, there’s an awful lot to like about the way the Beosound A1 3rd Gen goes about things. So minor are these caveats, in fact, I may as well get them out of the way right now before piling into all the things the Bang & Olufsen excels at.

It’s probably an inevitability given the physical size of the speaker, but the A1 3rd Gen isn’t an especially large-scale listen. The soundstage it creates is properly organised and quite open – but where out-and-out scale is concerned, there’s no getting around the fact that this speaker doesn’t sound especially big.

Its tonal balance isn’t, strictly speaking, all that neutral either. There’s a hint of warmth and richness, a kind of sonic luxuriousness, about its presentation that’s not in any way unpleasant but that is most definitely a characteristic. This can be mitigated a little in the control app – but given that the best overall balance comes from leaving the EQ setting at ‘optimal’ (the clue’s in the name, really) it’s probably best to accept your Bang & Olufsen speaker sounds, as well as looks and feels, deluxe.

From here on out, the news is all positive. The A1 3rd Gen can summon frankly unlikely low-frequency presence and substance - it hits remarkably hard for a speaker of its size, and the claim for frequency response down to 54Hz seems perfectly plausible. The low end is properly controlled, too, which means rhythms during Thin Air by Floating Points are confidently expressed – and the Bang & Olufsen is a front-foot listen as a result. At the opposite end of the frequency range it attacks with polite determination, bringing a reasonable amount of bite and shine to treble sounds without ever threatening to become unruly. There’s a fair amount of substance to offset the brilliance, too, which helps the overall balance of the sound no end.

In between, the A1 3rd Gen communicates through the midrange in the most uninhibited manner. Voices as dissimilar as Joni Mitchell’s and Alex Turner’s emerge with all of their character intact, and the midrange is nudged forwards just a little to help with intelligibility. The crossover between the mid/bass driver and the tweeter is smooth to the point of imperceptibility, and though the tonal balance here has a very definite flavour, it is at least consistent from top to bottom. The Bang & Olufsen sounds unified and of a whole, no matter the sort of content you’re listening to.

What else? Well, detail levels are high across the board – and even when it comes to the most minor harmonic variations or transient events in a recording, the A1 3rd Gen does good work in identifying and contextualising them. And despite its relative lack of outright scale, it can put appreciable distance between the quietest and the loudest moments in a recording too. As long as you’re not expecting to fill a medium-sized room with sound, the A1 3rd Gen is a great option – and you can at least be assured that if you do wind up the volume to the point the Bang & Olufsen is giving you everything it’s got, it doesn’t alter its sonic character in the slightest. It simply gets louder.

• Sound quality score: 5/5

(Image credit: Future) Bang & Olufsen Beosound A1 3rd Generation review: design

• 46 x 133 x 133mm, HxWxD
• 576g
• Built to a high standard from premium materials

As both the asking price and the manufacturer’s reputation demands, the Beosound A1 3rd Gen is little short of a masterclass where materials and their deployment are concerned. It’s tricky for a product of such modest dimensions (this speaker is just 46 x 133 x 133mm, HxWxD) to look overtly premium, but this Bang & Olufsen manages just fine.

The lower half of the spheroid is of hard-wearing polymer, and it’s seamlessly connected to an upper portion made of tactile, textured pearl-blasted anodised aluminium. A carry-strap made of waterproof leather is threaded through the polymer, and it’s here you’ll find the only visible bit of branding. The look is understated and sophisticated, no matter which of the three available finishes you choose – take your pick from ‘natural aluminium’ (which is ‘silver’), ‘honey tone’ (‘gold’) and ‘eucalyptus green’ (‘green’). And the 576g weight seems purposefully judged to strike a balance between ‘portability’ and ‘reassuring heft’. The overall standard of construction and finish is basically flawless, and the IP67 rating means it will be fine in any realistic environment.

This is, says the manufacturer, the first ever Bluetooth speaker with ‘Cradle-to-Cradle’ certification (at the bronze level). So it’s made with sustainable practices in mind, and is modular to the point that it has a replaceable battery and can, in certain circumstances, be repaired rather than scrapped. This is something Bang & Olufsen has long been known for; just see the company's modular Beosound Theater soundbar and reissued vertical CD player for starters…

• Design score: 5/5

(Image credit: Future) Bang & Olufsen Beosound A1 3rd Generation review: setup and usability

• Some physical controls
• Excellent B&O control app
• Swift Pair and Fast Pair

It’s a very simple device to operate, the A1 3rd Gen – the options you have are very tidily realised indeed. Along with that USB-C socket, the polymer edge of the chassis features a few physical controls, all very positive in their action – you can take care of ‘power on/off’, Bluetooth pairing, ‘volume up/down’, ‘play/pause’ this way. There’s also a button to mute the mics that are used for telephony. Microsoft Swift Pair and Google Fast Pair are available, while a single LED on the top surface indicates what the speaker might be up to, and another next to the USB-C input lets you know about charging.

There’s also the B&O control app that’s free for iOS and Android – and it remains one of the best examples of its type. Here’s where you can check for software updates, form that stereo pair using another A1 3rd Gen or your A1 2nd Gen, integrate your favourite music streaming service (as long as it’s Deezer, Spotify or TIDAL) and access music on a HomeMedia server. It also gives internet radio access via the entirely admirable ‘B&O Radio’. It’s where you can access some EQ presets, or define your own using a graphic ‘target’ control that’s about as gratifying an EQ adjustment mechanism as I’ve ever seen.

• Setup and usability score: 5/5

(Image credit: Bang & Olufsen) Bang & Olufsen Beosound A1 3rd Generation review: value

This is far from the most affordable Bluetooth speaker around – although when judged by the standards Bang & Olufsen has long since established, it’s actually within touching distance of many credible alternatives.

The usual B&O selling points of premium materials, indulgent design and impeccable build quality are all present and correct, too – and the Beosound A1 3rd Gen is a match for any price-comparable rival where pure sound quality is concerned. So for once here’s a Bang & Olufsen product that doesn’t require all that “yes, but..” when it comes to value for money.

• Value score: 5/5

(Image credit: Future) Should I buy the Bang & Olufsen Beosound A1 3rd Generation? Buy it if...

You like nice things
The Beosound A1 3rd Gen looks and feels every bit the Bang & Olufsen product.

You know a good control app when you see one
The B&O control app is comprehensive, easy to use and features the most gratifying EQ adjustment around.

You enjoy a hint of sonic luxuriousness
The A1 3rd Gen is slightly less than neutral in its sound, but in a quite expensive-sounding way.

Don't buy it if...

You intend to fill the great outdoors with sound
The A1 3rd Gen has a degree of grunt, but it’s not the biggest-sounding speaker around.

You want to just ask
There’s no Alexa integration for this latest model.

Bang & Olufsen Beosound A1 3rd Generation review: also consider

JBL Flip 7
The March 2025-issue seventh Flip is also a five-star Bluetooth portable speaker. And it also looks good, thanks to a new switchable carabiner detail. The battery won't last quite as long as the B&O Beosound A1 Gen 3, but the audio is special for a speaker of its size, and it's hard to ignore the price difference… 
Read just how good it is in our full JBL Flip 7 review.

Bose Soundlink Flex Gen 2
It's a September 2024 release, so still a relative newbie to the game, and although the Bose Soundlink Max is closer in price to the B&O option you came here to read about (at $399 / £399 / AU$599), the Flex Gen 2 is a better size match – and a lot cheaper. How often can you say that of Bose products? It's not quite as gifted sonically, but it does give you punchy output in a relatively bijou design. Get the full scoop in our Bose Soundlink Flex Gen 2 review.

Bang & Olufsen Beosound A1 3rd Generation review: how I tested

I connected the A1 3rd Gen wirelessly to my Apple iPhone 14 Pro and my FiiO M15S digital audio player, and also made a hard connection via USB-C to my Apple MacBook Pro. So I was able to listen to lots of different content from lots of different platforms, of various file-types and sizes.

And having done so for well over a week, I realised the A1 3rd Gen is the best yet – which is saying something…

• First reviewed: May 2025
• Read more about how we test

• Andor season 2 episode 9 finally sees Mon Mothma's infamous Senate speech brought to life
• Genevieve O'Reilly said she was "so nervous" about doing justice to the iconic moment
• Fans have waited years for it to be depicted in a live-action or animated project

Full spoilers immediately follow for Andor season 2, up to and including episode 9.

Genevieve O'Reilly has revealed why she was "so nervous" about finally delivering Mon Mothma's famous Senate speech in a Star Wars project.

Before Andor season 2 was released in late April, I sat down with O'Reilly to discuss the scene that's not only a landmark moment for the Chandrilan Senator but also for O'Reilly.

O'Reilly is renowned for portraying Mon Mothma in Andor on Disney+, but she's actually played the character in multiple projects since 2005's Star Wars Episode III: Revenge of the Sith.

Throughout her 20-year (and counting!) career as part of Lucasfilm's legendary sci-fi franchise, O'Reilly has never had the opportunity to bring her character's hugely significant speech to life before. That's changed now, though. To the delight of Star Wars fans who've waited years for this moment, Mon Mothma finally gets on her proverbial soapbox to denounce Emperor Palpatine for his autocratic rule of that iconic galaxy far, far away.

Benjamin Bratt's Bail Organa is the only Senator who agrees with Mon Mothma's stance on the Empire (Image credit: Lucasfilm/Disney+)

It's a hugely pivotal moment for Mon Mothma, the nascent Rebel Alliance, and the galaxy as a whole.

Indeed, while Mon Mothma has been part of Luthen's quickly evolving, galaxy-wide rebellion for years at this point, she's largely operated in the shadows to keep herself and her family safe from the Galactic Empire's prying eyes.

All of that is thrown out the window in this season's ninth episode. The emotionally devastating events in season 2 episode 8, which depicted the Ghorman Massacre in all of its harrowing glory – an event Andor season 2 has laid the groundwork for since its first episode – Mon Mothma decides to take a public stand against the Emperor and those who follow his every whim.

Mon Mothma's speech has been 20 years in the making for Genevieve O'Reilly (Image credit: Lucasfilm/Disney+)

Addressing the Senate in one of the best Disney+ shows' latest episodes, Mon Mothma not only condemns Palpatine for the slaughter of innocent and largely peaceful protestors on Ghorman, but outright rejects his tyrannical leadership.

Vilified by her fellow Senators (aside from Bail Organa, anyway, who silently supports her) and knowing that the Empire will come for her, Mon Mothma quickly flees the scene.

With the aid of Cassian Andor, who's been sent by Luthen Rael to get her to safety, Mon Mothma escapes Coruscant. Off-screen, she's passed onto the crew of The Ghost, aka the heroes of Star Wars: Rebels, in season 3 episode 18, titled 'Secret Cargo', of that animated program. A couple of Rebels characters appear in another live-action series Star Wars: Ahsoka, too, which further highlights the connective tissue between numerous Lucasfilm productions.

Now we know who delivers Mon Mothma to the crew of the Ghost in Rebels' 'Secret Cargo' episode (Image credit: Lucasfilm/Disney+)

But I digress. Andor has been building towards this significant Mon Mothma moment ever since we were reintroduced to her in mid-2022. O'Reilly has also harbored ambitions of giving this speech for almost two decades, too, so how did it feel to finally deliver it in one of 2025's new Star Wars shows?

"The opportunity to give that speech, as well as the lead up to it, was everything I could have hoped for," O'Reilly replied.

"I was so nervous that, because it's existed previously in canon, we might not get to visit it. But what [showrunner] Tony [Gilroy] did, which was so clever, is carve a space to allow us to see that moment.

I was so nervous that... we might not get to visit it

Genevieve O'Reilly

"I think it makes sense of everything about her," O'Reilly continued, "And not just her. It makes sense of our show. If sacrifice is at the cost of everyone, Mon has to understand and be at peace with what she has to sacrifice.

"I also think it's important to mention that a voice can be powerful. It's not about the fact that we pick up weapons, but perhaps a voice can shift things, and that speaks to everything that is why I think Mon Mothma is valuable to the rebellion."

Andor season 2 episodes 1 to 9 are out now on Disney+. The final three chapters will air on May 13 (US) and May 14 (UK and Australia).

You might also like

• 'I wanted to do justice': Andor season 2 star Adria Arjona on Bix's cathartic tale of revenge in the acclaimed Star Wars show
• Find out how to watch the Star Wars movies in order
• When might Marvel's Thunderbolts* movie come out on Disney+?

In today’s digital-first world, citizens naturally expect that public services should be as quick and easy to use as their smartphone apps.

In Britain, we are at a turning point in our approach to artificial intelligence, particularly in the public sector. Governments are increasingly recognising that upgrading the ‘citizen experience’ by investing in cutting-edge technology will be vital to delivering tomorrow’s public services.

Earlier this year, in the Government’s new ‘AI Opportunities Plan’, the Prime Minister highlighted how AI tools can “transform” public services and deliver seamless experiences similar to how we manage money or book flights online. Most recently, the Prime Minister’s announcement to “reshape” the state offers a huge opportunity to reinforce the UK’s ambitions to be an AI leader across the public sector.

AI adoption starts now

At ServiceNow, we are working with hundreds of public sector bodies and can see how AI is already helping to cut through challenges, such as the 13.5 million hours doctors lose annually to outdated IT. Virtual assistants and advanced triaging systems are now helping to cut through the ‘8am rush’ to book a GP appointment. We are seeing similar impacts in other departments such as the Driver and Vehicle Licensing Agency (DVLA) and Department for Work and Pensions (DWP).

For public services, the impact of technology is not some futuristic fantasy: it is here, now, today. The key to implementing this successfully is partnerships between the public and private sector. The Government’s plan calls for the public sector to become a “great customer” of AI services. And yet analysis by GDS suggests that almost half of Government IT spend is dedicated to ‘keeping the lights on’ maintaining legacy systems. In the private sector, businesses are implementing AI to drive productivity and enhance the customer experience. It is vital that the public sector does so too.

Delivering for citizens

The first step for government is replacing patchwork legacy systems with an AI platform to speed innovation, scale business processes, and forge a solid foundation. Analysis by the Social Market Foundation suggests that at HMRC and the DWP alone, eight million hours of staff time could be saved every year by using technology to streamline routine tasks. Moving away from legacy systems, siloed departments and patchwork IT is also crucial to delivering citizen experience, making it easier to connect to government services.

Rather than waiting for working-hours call centers and having to work out who the ‘right’ department is to speak to, digital channels and generative AI are offering real-time answers to citizens. Not only that, but AI’s capacity also to process large stores of data are making digital services more effective and accessible.

AI in public service

Forward-thinking government organizations have already adopted generative AI to deliver better service for citizens, as well as increasing efficiency and offering instant access to data. Large Language Models (LLMs) trained on domain-specific data are already helping government organizations respond rapidly, either by empowering agents with the information they need which might previously have been inaccessible, or by offering citizens direct access to data.

The future of government

From managing risk to delivering improved citizen experience, cutting-edge technology is the key to overcoming many of the hurdles seen in previous government attempts at digital transformation. Ripping out the siloed legacy systems, which are all too often seen in the public sector and replacing them with a single AI platform connecting all data across the organization, will not only drive efficiency, but also pave the way for more wholesale innovation.

The addition of AI, and of its latest evolution with Agentic AI, is a stepping stone towards a new era of innovative, automated services – with citizens able to access personalized experiences, where and when they want.

We've compiled a list of the best IT infrastructure management services.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

• Agentic AI adoption will rise by 327% by 2027, boosting productivity by 30%
• Workers will need to share their roles with AI agents soon, Salesforce claims
• Many workers could be reassigned roles, report says

AI agents are here to stay, with new Salesforce research claiming agentic AI adoption is projected to grow by 327% by 2027, with the company calling the trend a revolution of "digital labor."

Chief Human Resource Officers (CHROs) are expecting to keep 61% of their existing workforces in their current roles, however employees are set to be working alongside AI.

Most CHROs (88%) surveyed added the redeployment of human resources alongside tech can be more cost-effective than external hiring, suggesting workers' jobs could be more secure than they think, but that the change they're subjected to could also be greater.

Combining workers with AI

In line with the projected growth of AI agents, Salesforce believes an increase in productivity of 30% could be realized. The figures also forecast a 19% reduction in labor costs.

With AI literacy identified as the top skill needed in the modern workplace, four in five (81%) CHROs are reskilling or planning to reskill employees for future roles, including reassigning many to technical roles like data scientists and technical architects.

Among the teams expected to see the biggest growth are IT, research & development, and sales. Customer service, operations and finance are expected to shrink.

Fortunately, workers seem to have plenty of time to get their affairs in order and to embark on their training journeys, because 85% of organizations have not yet implemented agentic AI.

Unprepared workers don't have unlimited time, though, because 86% of CHROs believe integrating AI will be a critical part of their role within the next five years, with four in five believing that AI agents and humans will coexist in this timeframe.

"Every industry must redesign jobs, reskill and redeploy talent – and every employee will need to learn new human, agent and business skills to thrive in the digital labor revolution," explained Salesforce Chief People Officer Nathalie Scardino.

You might also like

• Prepare for the future with the best online learning platforms
• We've listed the best job sites and best recruitment platforms
• Amazon says one business per second is adopting AI

WeightWatchers said Tuesday it is filing for Chapter 11 bankruptcy protection to eliminate $1.15 billion in debt and focus on its transition into a telehealth services provider.

(Image credit: Richard Drew)

Here are the answers for The New York Times Mini Crossword for May 7.

Hundreds of flight delays and cancellations at Newark's airport over the past several days are giving passengers headaches and spurring promises to improve the nation's air traffic control system.

(Image credit: Spencer Platt)

• OpenAI’s latest AI models, GPT o3 and o4-mini, hallucinate significantly more often than their predecessors
• The increased complexity of the models may be leading to more confident inaccuracies
• The high error rates raise concerns about AI reliability in real-world applications

Brilliant but untrustworthy people are a staple of fiction (and history). The same correlation may apply to AI as well, based on an investigation by OpenAI and shared by The New York Times. Hallucinations, imaginary facts, and straight-up lies have been part of AI chatbots since they were created. Improvements to the models theoretically should reduce the frequency with which they appear.

OpenAI’s latest flagship models, GPT o3 and o4-mini, are meant to mimic human logic. Unlike their predecessors, which mainly focused on fluent text generation, OpenAI built GPT o3 and o4-mini to think things through step-by-step. OpenAI has boasted that o1 could match or exceed the performance of PhD students in chemistry, biology, and math. But OpenAI's report highlights some harrowing results for anyone who takes ChatGPT responses at face value.

OpenAI found that the GPT o3 model incorporated hallucinations in a third of a benchmark test involving public figures. That’s double the error rate of the earlier o1 model from last year. The more compact o4-mini model performed even worse, hallucinating on 48% of similar tasks.

When tested on more general knowledge questions for the SimpleQA benchmark, hallucinations mushroomed to 51% of the responses for o3 and 79% for o4-mini. That’s not just a little noise in the system; that’s a full-blown identity crisis. You’d think something marketed as a reasoning system would at least double-check its own logic before fabricating an answer, but it's simply not the case.

One theory making the rounds in the AI research community is that the more reasoning a model tries to do, the more chances it has to go off the rails. Unlike simpler models that stick to high-confidence predictions, reasoning models venture into territory where they must evaluate multiple possible paths, connect disparate facts, and essentially improvise. And improvising around facts is also known as making things up.

Fictional functioning

Correlation is not causation, and OpenAI told the Times that the increase in hallucinations might not be because reasoning models are inherently worse. Instead, they could simply be more verbose and adventurous in their answers. Because the new models aren't just repeating predictable facts but speculating about possibilities, the line between theory and fabricated fact can get blurry for the AI. Unfortunately, some of those possibilities happen to be entirely unmoored from reality.

Still, more hallucinations are the opposite of what OpenAI or its rivals like Google and Anthropic want from their most advanced models. Calling AI chatbots assistants and copilots implies they’ll be helpful, not hazardous. Lawyers have already gotten in trouble for using ChatGPT and not noticing imaginary court citations; who knows how many such errors have caused problems in less high-stakes circumstances?

The opportunities for a hallucination to cause a problem for a user are rapidly expanding as AI systems start rolling out in classrooms, offices, hospitals, and government agencies. Sophisticated AI might help draft job applications, resolve billing issues, or analyze spreadsheets, but the paradox is that the more useful AI becomes, the less room there is for error.

You can’t claim to save people time and effort if they have to spend just as long double-checking everything you say. Not that these models aren’t impressive. GPT o3 has demonstrated some amazing feats of coding and logic. It can even outperform many humans in some ways. The problem is that the moment it decides that Abraham Lincoln hosted a podcast or that water boils at 80°F, the illusion of reliability shatters.

Until those issues are resolved, you should take any response from an AI model with a heaping spoonful of salt. Sometimes, ChatGPT is a bit like that annoying guy in far too many meetings we've all attended; brimming with confidence in utter nonsense.

You might also like

• ChatGPT crosses a new AI threshold by beating the Turing test
• AI took a huge leap in IQ, and now a quarter of Gen Z thinks AI is conscious
• ChatGPT model matchup - I pitted OpenAI's o3, o4-mini, GPT-4o, and GPT-4.5 AI models against each other and the results surprised me

Hints and answers for the NYT Connections: Sports Edition puzzle, No. 226, for May 7.

• Gigabyte AI TOP 100 Z890 PC blends elite gaming hardware with AI-focused design and software
• AI TOP Utility supports LLMs with 405B parameters and real-time model monitoring tools
• Eight SATA ports and two NVMe drives make storage scaling effortless for creators and devs

Gigabyte has revealed a new desktop PC designed for users working with artificial intelligence in professional creative tasks and high-end gaming.

The company says its the system is equipped with Nvidia’s flagship GeForce RTX 5090 GPU and Intel’s top-tier Core Ultra 9 285K CPU, and introduces the new Z890 AI TOP platform.

The GeForce RTX 5090 WINDFORCE GPU includes support for DLSS 4.0 and Multi Frame Generation, making it suitable for machine learning workloads.

Advanced memory and cooling design

The Gigabyte AI TOP 100 Z890 business PC comes with 128GB of DDR5 memory, supported by D5 Bionic Corsa technology, which enhances stability during intensive AI and gaming sessions.

A unique feature is the inclusion of a 320GB AI TOP 100E cache SSD alongside a 2TB AORUS Gen4 7300 SSD. The latter is built to endure heavy data writes, boasting a lifespan up to 150 times greater TBW than standard SSDs.

The system also supports up to eight additional SATA drives, making it a viable option for users needing scalable storage without sacrificing speed.

To cool all this hardware, Gigabyte installed its AORUS WATERFORCE II 360 liquid cooling system, which keeps thermal output in check even under prolonged workloads.

While not a portable machine, the system maintains a standard ATX form factor, measuring 594 x 336 x 584 mm, and supports both Windows 11 Pro and Linux.

For connectivity, the workstation offers a solid I/O setup, including front-mounted USB 3.2 Gen 2 Type-C, USB 3.0, and audio jacks, as well as a rear panel with two Thunderbolt 5 ports, six USB 3.2 Gen 2 ports, HDMI, DisplayPort In, dual 10GbE RJ-45, Wi-Fi 7, and Bluetooth 5.3.

Gigabyte’s AI TOP Utility suite simplifies AI workflow management, supporting memory offloading and fine-tuning for models up to 405 billion parameters. It also includes tools for dataset creation, real-time training dashboards, and model validation. Cluster computing is enabled via Thunderbolt and Ethernet links, allowing scaling across multiple systems.

According to Gigabyte, “users can access AI computing with a standard power setup and plug-and-play convenience.” This new system targets a wide range of users - from researchers and developers to content creators and gamers.

At the time of writing, there is no official word on pricing or availability, but it’s not expected to be cheap.

You might also like

• These are the best business laptops available to buy right now
• And you should take a look at the best office chairs we've tried
• World Password Day 2025: All the news, updates and advice from our experts

• Google’s Gemini can now edit both AI-generated and personal images using text prompts
• The editing tools allow for precision changes using AI
• Gemini also now supports uploading up to 10 images or files at once

Google's Gemini AI is taking out a canvas and palette for your AI-fueled image creation in a couple of major upgrades. Gemini can now edit images directly within its chat interface, and you can send a bunch of images (or other files) for it to examine at the same time.

The new editor can work AI magic on any image you upload or that Gemini produces. You simply ask Gemini to make the changes you want. You can change the backdrop of your vacation photo to put your sad Airbnb kitchen on a Santorini cliff, get rid of that mustard stain on your jacket, and even put a funny hat on your dog despite her refusal of all headwear in real life.

You can apply multiple edits through your conversation with Gemini, stacking changes as you go. And each modification keeps earlier changes, so you don't have to start from scratch when you decide the last couple of edits made things worse.

Under the hood, Gemini’s editor is running a combination of tools that work together so you don't end up with a visual Frankenstein's monster stitching together conflicting textures, lighting, angles, and other aspects of the image. Gemini promises to keep things grounded in reality even when your imagination goes off the rails.

Google claims the editor will have many positive uses for a range of professions. Teachers could quickly build illustrated storyboards, designers could make a portfolio of product photos, and architects might visualize tweaks to building designs mid-meeting.

The editor pairs nicely with Google's move to blow up the single-file upload limit for Gemini. Now you can upload up to ten images, PDFs, or other files all at once and ask Gemini to make sense of the mess.

AI image imagination

You may be wondering how Gemini's editor will prevent people from leveraging its abilities to make deepfakes of real people or events for less than benign reasons. Google is keen to show that the company has thought of that. That's why every AI-edited image gets not one but two watermarks. One is visible, and one uses Google’s SynthID, which can only be detected with software. There are also filters powered by human feedback that block ethically dicey requests.

The editor and expanded upload option are not breaking new ground, but they add depth to Gemini. It’s not just about what Gemini can tell you, it’s about what it can help you make. Google is investing a lot of effort in building Gemini into the kind of well-rounded, versatile toolkit that people are comfortable relying on.

Instead of thinking of Gemini as a mere digital notetaker or search engine with a sense of humor, Google wants people to view Gemini as a partner in creative and productive tasks. We’re still a ways off from a world where you ask Gemini to “design a birthday card and bake the cake,” but it's closer than you might think. Until then, being able to throw ten files at Gemini and have it respond with something coherent while also placing a hat on your dog is a pretty good start.

You might also like

• I tried Gemini's new AI image generation tool - here are 5 ways to get the best art from Google's upcoming Flash 2.0 built-in image upgrade
• I compared Adobe’s new Firefly Image Model 4 to ChatGPT’s image generator, and it’s like they went to the same art school
• I turned my dog into a plushie using AI and it was super easy

Between tariffs roiling the global economy and OPEC putting more barrels on the market, prices have dropped significantly since January. That's good for consumers, and bad for oil companies.

(Image credit: Andrew Harnik)