Feed aggregator

• Bitdefender found more than 300 Android apps used to display unwanted ads
• Almost all apps have so far been removed from the Play Store
• Some of the apps even tried to steal sensitive data

Hundreds of Android applications, installed millions of times from the Google Play Store, were part of a large-scale ad fraud campaign that displayed unwanted ads and tried to steal sensitive data from the victims, experts have warned.

Cybersecurity researchers from Bitdefender, who also credited IAS Threat Lab, said they discovered at least 331 apps that were available via the Google Play Store which, cumulatively, amassed more than 60 million downloads, mostly on older Android variants (Android 13 and older).

The apps were mimicking simple utility apps such as QR scanners, expense tracking apps, health apps, wallpaper apps, and others.

Protecting your Android phone

Most applications first became active on Google Play in Q3 2024, and by the time Bitdefender’s research completed, only 15 remained active. The vast majority of victims are located in Brazil, followed by the US, Mexico, Turkey, and South Korea.

When the victims downloaded one of the apps, they first hide their icons from the launcher (something that’s only possible on older versions of Android). The apps retained “some functionality”, but they are able to show out-of-context ads over other applications in the foreground. Some try to gather user credentials, credit card data, and other information.

Some apps can even start without user interaction, something that shouldn’t be possible even on Android 13.

While eventually all of the apps will be removed from the Play Store, people who have them installed on their devices remain at risk. If your Android smartphone is acting strange (lagging, displaying ads over different ads, overheating, spending data while idle), make sure to remove any unwanted apps, or those that you aren’t actively using.

Furthermore, if possible, make sure to use the latest version of Android. At the moment, that is Android 15, with the next iteration scheduled for later in 2025.

You might also like

• Windows PCs targeted by dangerous new threat that even gets around Defender - and even though there's a fix, you could still be at risk
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

• Microsoft is sending out emails to push people to upgrade from Windows 10 to Windows 11
• While on the face of it, that seems a useful move to help some users, Microsoft’s angling of the email is far from ideal
• It also runs the risk of making Windows 10 users feel spammed, particularly as they’re still getting nudged numerous times within the OS itself

Microsoft is once again trying to persuade Windows 10 users that they need to upgrade to Windows 11, ahead of the impending cessation of support for the older operating system later this year.

This time, though, the nudge to upgrade isn’t being delivered within Windows 10 itself, but via email – although it isn’t the first time Microsoft has tried this approach.

I received an email from Microsoft (sent to the email address linked to my Microsoft account) regarding my Windows 10 PC needing an upgrade at the end of November 2024, a few months back, but now the software giant is sending out fresh messages to upgrade this month.

I didn’t get this latest mail (not yet, anyway), but Windows Latest did, and although it carries the same title, a warning that ‘End of support for Windows 10 is approaching,’ the email itself is somewhat different.

The overall thrust of the content is similar though. There’s a prominent reminder of the exact date that Microsoft halts support for Windows 10 – which happens on October 14, 2025 – and some suggestions of what to do with your old PC (trade it in, or recycle the machine). You can also click a link to check your upgrade eligibility for Windows 11.

Microsoft also clarifies that your PC will continue to work, it’s just that there will be no more support – as in software updates – piped through. There’s also a link to some blurb on how Windows 11 is more secure (which is certainly true), and a nudge to use OneDrive to back up your files if you plan to use Windows 10 after the deadline has passed, heading into 2026.

(Image credit: Future / Jeremy Laird) Analysis: Overstepping boundaries and taking some odd angles

There are a couple of things that strike me as odd here. Firstly, the plug for OneDrive feels very gratuitous, and hardly a solution to counter the prospect of having your PC compromised by running an out-of-date OS. Where on earth is the stern warning that it really isn’t a good idea to run Windows 10 on your PC when support for the operating system expires?

As you may be aware, without security updates, your computer will be left vulnerable to exploits, as when holes appear in Windows 10, they will no longer be patched up – a recipe for disaster, potentially.

Of course, if you really want to stick with Windows 10, then for the first time ever, consumers can pay to extend support, and I’d recommend you do so (for other options, explore my article on how to prepare for Windows 10 End of Life). Oddly enough, Microsoft doesn’t mention this extension of support in its email.

I say it’s odd, but then, Microsoft would really prefer you upgrade to Windows 11 anyway, either on your current PC – if it’s eligible – or by purchasing a new Windows 11 computer. And to that end, there’s a link in the email to ‘explore new computers’ which is something Microsoft has been urging us to do for a while now. As I’ve discussed before, there’s arguably merit to the suggestion in some ways, but a whole lot of other concerns outweighing that around the environmental toll that a ton of Windows 10 PCs ending up on the scrapheap might usher in.

These are serious worries, and likely why Microsoft is sending the other message in this email advising on recycling (or trading in) your old Windows 10 PC if you do upgrade.

The other point here is do you want to be getting emails direct from Microsoft about Windows 10 upgrades? Well, in some ways, I guess it’s better (or at least slightly less annoying) than being pushed to upgrade within the operating system itself, but the problem is, Microsoft is doing that as well – so Windows 10 users are getting both barrels, as it were. Sigh…

We can likely expect several further barrages of these kind of emails as 2025 progresses, and the October support deadline draws nearer – messages that folks may well be wanting their spam filter to deal with, frankly.

Don’t get me wrong here: I’m not saying it isn’t important to warn consumers about the dangers of an out-of-date operating system – it definitely is – but Microsoft is rather overstepping with its broad approach here, and worse still, this particular email actually undersells those dangers (while overselling other Microsoft products).

You may also like...

• Shock, horror – I’m not going to argue with Microsoft’s latest bit of nagging in Windows 11, as this pop-up is justified
• Microsoft is supercharging Windows 11’s voice commands on Copilot+ PCs with Snapdragon CPUs, and fine-tuning a few Recall features
• 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

This month marks five years since the pandemic began, and here are 5 things that changed permanently.

(Image credit: Dr. Kurt Papenfus)

• Adobe launches 10 new CX and marketing agents at Summit 2025
• One of them will optimize your website for you completely hands-free
• AI agents are also enhancing customer interactions

Adobe has introduced 10 new purpose-built AI agents it says can enhance customer experiences and marketing workflows across its Adobe Experience Cloud and has already promised to add more soon.

Unveiling the new offerings at its Adobe Summit 2025 event, the company boasted of the one trillion experiences per year its Adobe Experience Platform activates, which gives its latest AI tools a “deep, semantic understanding of enterprise data, content and customer journeys.”

Adobe’s approach revolves around simplicity, and while many enterprises might want to personalize their agents, the pre-configured tools have data governance and regulatory compliance built-in off-the-shelf, which all helps companies get up to speed more quickly.

Adobe agents - and an Agent Orchestrator

To manage agentic AI, Adobe has launched a new Adobe Experience Platform Agent Orchestrator interface, which works with Adobe’s own agents and third-party ecosystems.

From launch, the 10 new agents include: account qualification, audience, content production, data insights, data engineering, experimentation, journey, product advisor, site optimization and workflow optimization agents.

Adobe’s content production agent is perhaps a standout, allowing marketers and creatives to automate parts of their workflows by generating and assembling content based on a brief. The system adheres to pre-defined brand guidelines to produce accurate results the first time around.

The site optimization agent is also noteworthy for its always-on support for performant brand websites – it can automatically detect, recommend and then fix issues to boost customer engagement.

“Adobe is uniquely positioned to help brands meet this moment, with deep expertise in unifying AI, data and content production workflows to execute the right digital experiences with precision, while uncovering unseen problems,” said Digital Experience Business SVP Amit Ahuja.

At the same time, Adobe launched its Brand Concierge, a new AI application that enables conversational experiences tailored to individual brand interactions by using a company’s unique brand attributes and customer data.

Citing a 1,200% increase in traffic to US retail sites from GenAI sources between July 2024 and February 2025, Adobe Brand Concierge can create a more tailored experience in B2C settings, for example booking follow-up meetings on behalf of customers, but it also has B2B applications such as connecting sales and marketing teams.

Explore the power of generative AI with Adobe Firefly

Part of Adobe's growing AI suite, Firefly is tailor-made for teams who need business-critical, on-brand AI generated custom content. Want to see how Adobe's pro-level generative AI can help your business? Try out Firefly's tools by clicking here.

You might also like

• Craft your best work with the help of the best AI writers
• Check out the best customer experience tools
• Developers are pretty optimistic about AI agents

This review first appeared in issue 353 of PC Pro.

Workstation technology is the pinnacle of desktop PC power, and this month we’re testing a computer that showcases some of the latest high-performance components. Not only is the Scan 3XS GWP TR Ada our first look at AMD’s latest Ryzen Threadripper, back after a hiatus, but also Nvidia’s Ada generation graphics, and it boasts a bevy of PCI-E 5 NVMe storage as well.

First, that processor. The AMD Ryzen Threadripper 7980X is top of the resurgent range, offering 64 cores running at a base 3.2GHz but with a boost up to 5.1GHz. Scan has chosen Corsair H150i Elite Capellix RGB liquid cooling to eke the most out of this processor. The 360mm version has been installed on the front of the chassis, with three 120mm fans providing plenty of heat dissipation.

With the Ryzen Threadripper supporting quad-channel memory, Scan naturally populates all four DIMM sockets of the Asus Pro WS TRX50 Sage WiFi motherboard. These are 32GB 4,800MHz Samsung ECC Registered DDR5 modules, for a total of 128GB. You would need to replace all of them if you wanted more memory up to the motherboard’s 1TB maximum, but 128GB should be enough for the lifetime of this system.

The powerful components are kept cool by Corsair’s H150i Elite Capellix RGB liquid cooling (Image credit: Future)

No expense spared

Another top-level component is the Nvidia RTX 6000 Ada graphics, which uses Nvidia’s Ada Lovelace architecture, a 4nm design that takes professional GPU performance to the next level and is used in the incredibly potent GeForce RTX 4090 consumer graphics card. The RTX 6000 Ada packs 18,176 CUDA cores, a considerable upgrade on the 10,752 cores of the A6000 it replaces. Both cards have 48GB of GDDR6 memory, but the Ada’s memory is faster (2,500MHz versus 2,000MHz) so bandwidth has increased from 768GB/sec to 960GB/sec. The GPU clock is also up to 2,505MHz on boost, compared to 1,800MHz, so this is a seriously fast 3D graphics accelerator.

Scan has spared no expense when it comes to storage, either. The boot drive is a single 2TB Corsair MP700 NVMe M.2 device. This takes advantage of the AMD Ryzen Threadripper 7000 series’ support for PCI-E 5. Running CrystalDiskMark 8, it delivered 11,492MB/sec reading and 11,773MB/sec writing, which are phenomenal levels of throughput. As if this wasn’t enough, there’s an 8TB array consisting of four more of the same drive in a RAID0 configuration, installed in a Hyper M.2 X16 expansion card that sits in one of the motherboard’s x16 PCI-E 5 slots. This array delivered mind-blowing 34,186MB/sec reading and 42,176MB/sec writing. That’s not a misprint – you’re getting 30 to 40GB/sec throughput. This will make editing even 8K video relatively painless, for example.

The drives offer throughputs of incredible speeds (Image credit: Future)

Touch of style

This plethora of potent components has been housed in a sizeable Fractal Design North chassis. The slats down the front make this look more like a stylish freestanding radiator than a PC, and it’s packed with features. There are two mounts for 2.5in drives, and two more for 3.5in or 2.5in units. None are used as all of Scan’s storage is NVMe-based. The power supply is a 1,000W Corsair RM1000X 80 Plus Gold unit, which should be well within its comfort zone despite the power-hungry core components. Even under full load, this isn’t a noisy computer.

The Asus motherboard offers both 10GbE and 2.5GbE LAN on the backplane, plus a built-in Wi-Fi 7 2x2 radio that can operate at up to 2.8Gbits/sec. There are three PCI-E 5 slots, but the two x16 ones are taken by the graphics and RAID storage array adapter, while the third operates in x8 mode. There’s also an x16 PCI-E 4 slot with a second that operates in x4 mode. The board contains three M.2 storage slots, with two operating at PCI-E 5 x4 speed (one taken by the boot drive), plus a third with PCI-E 4 x4 performance. The motherboard also has hardware and software support for Asus IPMI remote management expansion cards. However, there are only two 20Gbits/sec USB-C ports – one on the backplane and one on the top of the chassis at the front. The rest are Type-A ports of lesser speeds.

There are 10GbE and 2.5GbE LAN ports on the back, plus Wi-Fi 7 within (Image credit: Future)

Blasted with speed

We were expecting top performance from the new Threadripper, but we weren’t quite ready for just how fast this system is. An overall score in the PC Pro benchmarks of 954 is way ahead of anything we’ve tested before. Most of this is provided by the incredible 1,224 in the multitasking test, which is no surprise with 64 cores available. But the video result of 892 is phenomenal as well. Only the image editing score of 267 has been beaten before. This is a mostly single-threaded test, and there are CPUs with faster boost core frequencies.

If you want a true demonstration of how incredible this system is, however, look no further than the Maxon Cinebench R23 multithreaded result of 107,189. Putting this in perspective, the fastest 64-core Threadripper Pro 5995WX workstation we’ve tested – the Armari Magnetar MC64TP – only managed 76,330. The single-threaded result of 1,728 shows the improvement, too, as the Magnetar managed 1,492, although CPUs with lower cores and higher frequencies such as the Ryzen 7000 series or recent Intel Core chips do better in this test.

Further underlining these abilities, the Blender Gooseberry render took only 94.6 seconds using the CPU, the first time we’ve seen a processor complete this frame processing task in less than two minutes. On GPU (using CUDA) it took 69.2 seconds, much less time than any other professional card we’ve tested. The LuxMark 3.1 score of 26,426 is similarly high, showing just how much GPU compute power the Nvidia RTX 6000 Ada has on offer.

Power trip: the 1,000W Corsair PSU is an 80 Plus Gold-certified unit (Image credit: Future)

Tour de force

This GPU power translates into incredible 3D modelling capabilities. The standout result in SPECviewperf 2020 is an astonishing 1,235 with the snx-04 viewport, almost twice the fastest scores we’ve seen before. The catia-06 result of 208 is also impressive, as is 604 in solidworks-07, showing that CAD and product design capabilities are unparalleled, although the creo-03 score of 76 is surprisingly mediocre. For 3D animation, 295 in 3dsmax-07 and 807 in maya-06 tell their own story.

The Scan 3XS GWP TR Ada is a tour de force. Its 64-core AMD Ryzen Threadripper blows everything that went before out of the water with multithreaded tasks, and it’s no slouch with single threads, either. The Nvidia RTX 6000 Ada graphics also dominate for viewport acceleration or GPU rendering, and storage throughput is unparalleled. Of course, this workstation is expensive, but you certainly get what you pay for: the best possible workstation performance on the market.

There are plenty of good reasons to welcome back the non-Pro Threadripper, which is better than ever (Image credit: Future)

For a few years, it seemed like the regular AMD Ryzen Threadripper was history. After the Pro version arrived, there was no non-Pro 5000 series. But now the vanilla Threadripper is back with the 7000 generation. It’s based on the same Zen 4 core as the Ryzen and Ryzen Threadripper Pro 7000 series CPUs, manufactured using variants of TSMC’s 5nm fabrication process, but sits between the two in features and price.

Where the Ryzen 7000 series tops out at 16 cores, the Ryzen Threadripper 7000 series offers 24-, 32- and 64-core iterations. The Pro goes from 12 to 96 cores. Further differentiation comes with memory configuration. All three 7000 series ranges support DDR5 memory at up to 5,200MHz, but where the basic Ryzen has a dual-channel controller, the Threadripper’s controller is quad-channel and the Pro’s is eight-channel.

All three ranges now offer PCI Express 5, but the number of lanes available is another point of difference. The basic Ryzen has 28 PCI-E 5 lanes, but the Threadripper has 48, plus 24 PCI-E 4 lanes, and the Pro has 128 PCI-E 5 lanes. The greater number of lanes will mean that faster storage devices such as NVMe M.2 drives can be attached, faster networking devices can be used, and more graphics cards can be installed.

Since we last had non-Pro Threadrippers, AMD’s Zen core technology has moved on apace. Both the previous 3000 series and new 7000 series top out at 64 cores, but the 3000 version (3990X) had a base clock of 2.9GHz with 4.3GHz boost, whereas the 7980X as found in this month’s Scan system has a base clock of 3.2GHz with 5.1GHz boost. This will offer significantly better performance with software that favors fewer, faster cores and with applications that can use as many cores as possible. With AMD’s Precision Boost 2, you’ll get as many cores as your cooling and motherboard power delivery can handle to run at maximum frequency, although some will be flagged as the most capable after manufacturer testing. Putting this further in perspective, the 32-core AMD Threadripper 2990WX only had a base clock of 3GHz and a boost of 4.2GHz, so the 7980X is a huge leap forward.

As such, this Zen 4-based Threadripper represents another significant update for AMD. Intel has been posing a challenge to the Ryzen 7000 series with its 13th generation Raptor Lake chips, but its Sapphire Rapids Xeon workstation processors have only arrived recently and max out at 56 cores and a 4.8GHz boost frequency. The Threadripper undercuts it in price while the Threadripper Pro now goes up to 96 cores.

With the return of the non-Pro Ryzen Threadripper, there’s a more affordable route once again if you want more than 16 cores in your workstation for the best possible multithreaded performance.

We've also ranked the best large SSDs and hard drives.

This review first appeared in issue 353 of PC Pro.

As the latest member of the HP Poly Studio family, the X52 video bar delivers an advanced all-in-one solution for medium-sized meeting rooms. Taking over from the Studio X50, it features an uprated 20MP 4K UHD camera, a more powerful Qualcomm 865 CPU, improved dynamic framing algorithms, advanced microphones and Wi-Fi 6 support.

Apart from the larger camera there’s little to tell the two models apart but at 95°, the X52 has a narrower FoV (field of view) than the X50. This is because the X52 is designed to deliver a higher resolution when zooming in to participants at the far end of the meeting table.

Its mic array claims a six-meter pick-up range, while sound output is handled by two 20W ported speakers. You have two HDMI-out ports for dual monitor support, HDMI-in and USB-C for host connections, a gigabit network port and a port for Poly’s optional extension mic. All are easily accessible at the rear.

The X52 runs Poly’s Android-based VideoOS software, which supports BYOD (bring your own device) mode and includes built-in video apps. The guest list is impressive, and includes Microsoft Teams, Zoom Rooms, Google Meet, GoToRoom, RingCentral Rooms and BlueJeans Rooms.

Deployment is simple: you connect power and network, wait a minute while it boots up and point a browser at its IP address. The web interface presents a system status dashboard with the active video app and you move to the Provider section to choose another app from the drop-down list.

You can swap between them as required and, although a reboot is necessary, it only takes two minutes to regain access. BYOD mode is always available with the X52 automatically swapping to it when it senses a USB connection and, once you’ve finished, you can exit back to the app.

The Studio X52 delivers excellent video quality and can be remotely managed (Image credit: Future)

We had no problems cloud-managing the X52 by registering it with our Lens portal account. You can remotely manage all your Poly devices, assign them to sites and rooms, run remote firmware updates and upgrade to the premium Lens service for detailed meeting analytics.

We tested with Microsoft Teams and used the registration code presented on the screen to assign it to our Microsoft 365 account, where the X52 appeared as a new Teams Rooms on Android device. We also paired the X52 with Poly’s TC10 PoE-powered controller tablet and used the same registration process.

Our Teams console showed both devices as online and paired, after which the tablet moved all meeting controls to its own display. We had no problems creating and joining Teams meetings using the tablet’s 10in touchscreen, and in BYOD mode we could use its menu to manually control the camera and create up to ten presets for specific camera positions.

Video quality is excellent, and we noted the camera’s focus is markedly sharper than the Poly P15 we use in the lab. The mics worked well, as remote participants could hear us clearly when we were 4.5 meters away, while the dual speakers delivered a clean sound quality that easily filled our 24m2 meeting room at a 60% volume level.

The X52 gets the benefit of Poly’s NoiseBlockAI and Acoustic Fence technologies, which remove annoying background noises, while the DirectorAI feature handles group and speaker tracking. Also accessed from its web console, the DirectorAI Perimeter option allows you to define a custom tracking zone.

Although still in a preview testing phase, the X52 offers people framing, which shows all participants using up to six split screens. Whichever mode you choose, you’ll be impressed with its speed as the camera snaps to active speakers in no more than two seconds.

Poly’s Studio X52 is perfect for businesses that want a professional video conferencing solution for their medium-sized meeting room. Video quality is excellent, speaker tracking is impressively fast and the big choice of built-in VC apps makes it incredibly versatile.

We've also rated the best small business routers.

This review first appeared in issue 353 of PC Pro.

Despite its competitive price – the range starts at £649 inc VAT for a Core i3-1315U/8GB/256GB spec with Windows 11 Home – the E14 Gen 5 is everything a ThinkPad is expected to be. This graphite black laptop has the chrome ThinkPad logo on its aluminum lid, complete with the red-lit dotted “i” that gives it a hint of business elegance.

I tested the version with a polycarbonate/ABS cover, but you can choose aluminum. This adds nothing to the price and only 20g to the weight, and the latter is one of this laptop’s most attractive features: 1.4kg with the 47Wh battery, 1.5kg with the 57Wh option. However, with the bigger battery in place, the laptop only survived 7hrs 48mins in our web-surfing test. With this spec, we expected something over nine hours.

The power drain doesn’t come from Lenovo squeezing every last drop out of the supplied Core i7-1355U CPU. Although the E14 Gen 5 never missed a beat in my daily usage, the benchmarks don’t lie: Lenovo’s own ThinkPad X1 Carbon Gen 11 scored 8,979 in Geekbench 6.2’s multicore test to the E14’s 8,000. Similarly, in our Handbrake test, the E14 took 10mins 57secs to convert our sample 4K video to 1080p compared to 9mins 5secs for the X1.

(Image credit: Future)

Both machines have 16GB of RAM and, while the X1 has the advantage of LPDDR5 memory, it’s all soldered onto the motherboard. Here, 8GB is soldered on but there’s an 8GB DIMM that can be replaced by a 32GB module. You can also swap out the supplied 512GB SSD, a short but speedy M.2 2242 unit, or specify a 1TB model for an extra £60 when ordering.

The E14 continues Lenovo’s tradition of excellent keyboards, with snappy, deep-travel, chiclet-style keys, the trademark red TrackPoint and plenty of spacing. I loved typing on it and managed 91 words per minute in the 10FastFingers.com test with 97% accuracy, compared to my average of 88wpm with 91% accuracy. A smooth and centrally located touchpad adds to its attractions.

As ever with budget laptops, Lenovo has made a sacrifice with the screen. This is no wide-gamut affair, covering 65% of the sRGB space and 46% of DCI-P3. However, it’s strong in other areas: peak brightness of 374cd/m2 is actually 74cd/m2 more than Lenovo claims, while an average Delta E of 0.3 confirms excellent colour accuracy. It’s anti-glare, too, so great for office work, while a 1,920 x 1,200 resolution looks sharp over the 14in diagonal.

The excellent keyboard makes speedy typing for long periods a breeze (Image credit: Future)

A webcam perches above the screen, and it’s worth choosing the 1080p unit in my machine over the 720p option (at a cost of £20). It produced excellent video in good lighting, and even coped in low-light conditions. Autofocus and tracking were speedy, while using Windows Hello proved a breeze.

The dual mics did a fine job of focusing on my voice while keeping background noise to a minimum, and I was similarly pleased with the speakers.

Music emerged with discernible depth, mid-tones and highs. Take Megan Thee Stallion’s “Her”. The house music bassline hit hard, with no distortion even at the highest volume. Despite being lightweight and pretty slim – 19mm at the front, 22.7mm at the rear – Lenovo includes a generous number of ports. The left side packs two Thunderbolt 4 ports, a USB-A port, an HDMI out and a combo audio jack. The right features a USB-A port, gigabit Ethernet and a nano Kensington lock. The latter adds to an impressive roster of security features, with a physical cover on the webcam, a self-healing BIOS and the option of a fingerprint reader (£10) built into the power button.

Ethernet and plentiful USB ports make this a fine workhorse (Image credit: Future)

Our price includes all these options, including Windows 11 Pro, but one of the great things about Lenovo’s approach is that you can customize it. For example, switching to a Core i5-1335U saves £200. You might prefer to spend some of this to extend the one year of on-site cover to three years; this has a list price of £210, but Lenovo frequently offers generous discounts.

So while battery life disappoints and we have criticisms of the screen and its outright speed, the E14 Gen 5 adds up to a strong choice for businesses on a budget. And even if you do cut it down to its most basic specs, it will exude professionalism for years to come.

We've also ranked the best video editing software for beginners.

• The Xbox Adaptive Joystick is now available to purchase from the Microsoft Store
• It's fully compatible with the Xbox Adaptive Controller and other Xbox devices
• It can be bought now for $29.99 / £24.99

Microsoft has officially launched the Xbox Adaptive Joystick, which can be purchased now exclusively at the Microsoft Store.

Available for $29.99 / £24.99, the Xbox Adaptive Joystick is a wired Xbox accessory that connects via USB-C to other Xbox devices, such as an Xbox Series X or Series S console or Xbox Wireless Controller. It's also compatible, and perhaps best used, with the Xbox Adaptive Controller.

The Xbox Adaptive Joystick is quite sleek and simple in design. Its flat top features a single analog stick that can be pushed in similar to those on the Xbox Wireless Controller. The front of the joystick houses one bumper and one trigger-style buttons, while four face buttons rest at the top-middle of the device.

Players are able to fully customize their button layout on the Xbox Adaptive Joystick. They can also create and swap between multiple button profiles via the Xbox Accessories app which is available on both console and PC.

It also seems like the Xbox Adaptive Controller will be a fuss-free setup too. As a Windows device, it'll seamlessly work with Xbox Series consoles and Windows PCs without any updates required.

Microsoft's accessibility push on Xbox has been quite impressive, with dedicated controllers available to meet the needs of disabled players. Sony has also thrown its accessibility hat into the ring with the PlayStation Access controller, which our own managing editor, Rob Dwiar, awarded four out of five stars in his review.

You might also like...

• This SteamOS update promises a new future for non-Steam Deck handheld PCs – and I can’t wait
• Sennheiser announces new HD 550 headphones with high-quality audio for gamers and audiophiles
• This new HyperX wireless gaming headset can last for up to 200 hours before running out of juice

This review first appeared in issue 353 of PC Pro.

Kyocera’s EcoSys P2235dn is a compact mono laser printer, designed for moderate use in small groups or businesses. It’s well specified for the price, with a 250-sheet cassette and 100-sheet multipurpose feed, plus duplex printing as standard. It’s also equipped with a gigabit Ethernet port, but there’s no Wi-Fi support; you’ll need the P2235dw for that.

Most laser manufacturers regard photoconductor drums as a consumable, typically good for 10,000 to 40,000 pages, but Kyocera fits a ceramic drum designed to last the life of the printer; in this case, 100,000 pages. The drum is guaranteed to reach that, or to last for three years, but in the UK the printer itself gets a stingy one-year RTB warranty.

The P2235dn has a basic control panel comprising a few lights and buttons, so it’s odd to see it also has a USB host port, which is typically paired with a text or touchscreen so you can browse and select which contents to print. We searched Kyocera’s website and documentation, experimented with different files and button combinations, checked in the driver and combed through the web admin page, but simply couldn’t find out how to make direct prints.

Kyocera’s ceramic drum is designed to last the life of the printer (Image credit: Future)

Using our trusty PC, we subjected the P2235dn to our usual battery of print jobs. It’s a fast device, turning out a first page of text in 17 seconds, and continuing at 28.8ppm over our 25-page job. It extended this to 29.4ppm on a 50-page document, and even reached 12.5ipm when duplex printing ten sides of graphics onto five pages. We were particularly impressed by its 27.7ppm on our challenging 24-page graphics test.

Sadly, the P2235dn seems in too much of a hurry to print particularly well. It’s rare that we criticize text quality from a laser, but up close, character outlines weren’t as crisp as we’d expect. More significantly, mono graphics suffered from fine banding, something we’d usually only expect on the cheapest laser devices.

When considered with its 1.8p per page running costs, this lackluster print quality takes the shine off the P2235dn. It’s a good choice if speed is all important, but there’s little else to recommend it.

We've also rated the best video conferencing software.

• Nvidia is collaborating with Google and Disney to create a physics engine for robotics
• The open-source engine is titled Newton and is expected to launch later this year
• Nvidia CEO Jensen Huang revealed Blue, a Star Wars-inspired robot using Newton for complete real-time simulation

Nvidia CEO Jensen Huang has announced a new collaboration between the company, Google DeepMind, and Disney Research which is bringing AI-powered Star Wars robots to life.

Taking to the stage at Nvidia's GTC 2025 keynote on Tuesday, Huang revealed Blue, a Star Wars-inspired research robot capable of incredible movement akin to those seen in your favorite sci-fi movies.

The companies have teamed up to create Newton, the physics engine behind the robot's movement, which is expected to be released as open source later this year.

Huang said, "Can you believe you're looking at complete real-time simulation? This is how we're going to train robots in the future. Blue has two Nvidia computers inside." Nvidia's CEO went on to interact with Blue on stage before telling the robot to go home.

Nvidia's press release reads, "Newton is open source, empowering the entire robotics community. This enables roboticists to use and distribute the framework freely and contribute cutting-edge research to its development."

Now, this is all very proof of concept for the consumer, so what does Nvidia, Google, and Disney's collaboration mean for you and me? Well, we might not reap the rewards any time soon but after seeing Blue in action on stage, I'm convinced Disney's dream of having droids in Disney World is now going to become a reality.

In fact, just last week a report from Axios at SXSW stated that Disney is planning to showcase the robots in its entertainment parks at some point this year.

I've always wanted a robot, please make this a reality

You know, robots are pretty scary, I get it. But as someone who constantly struggles to deal with the stress of life in the 21st century, I'd absolutely jump at the opportunity to have a Star Wars droid in my home.

While I love my French Bulldog, Kermit, he can't do the dishes, he can't do the washing up, and he sure as heck can't understand what I'm saying (although I think he chooses to ignore me).

Now, this concept of a small cute robot doing all my chores in my home is not going to arrive anytime soon, but Newton's physics engine makes it a real possibility in the near future, and I'm sold on that idea.

Give me a robot that can make my life easier through the power of AI and I'll take out a loan to get one. My productivity would soar, my mental health would improve, and best of all, I'd hopefully never have to wash the dishes again.

You might also like

• “The age of generalist robotics is here" - Nvidia's latest GROOT AI model just took us another step closer to fully humanoid robots
• Will 2025 be humanoid robotics' breakout year? ChatGPT and Gemini have some fascinating thoughts
• Robot abuse is still the path to the future – even if I hate it

Full spoilers follow for Daredevil: Born Again's first four episodes and Echo.

We're halfway through – well, give or take – Daredevil: Born Again's first season. The arrival of episode 4, titled 'Sic Semper Systema', on Disney+ pretty much marks the midway point of the show's debut outing. And, thankfully, Marvel is finally starting to peel back the layers on some of its biggest mysteries.

So, if you want more details on Adam and/or Muse, the meaning behind this episode's title, or answers to an entirely different question, I've got you covered. Major spoilers follow for Daredevil's standalone Marvel Cinematic Universe (MCU) adventure, so proceed with caution if you're not caught up.

Who is Daredevil: Born Again villain Muse and who is playing him?

Muse's backstory might be expanded upon in the MCU compared to his comic book counterpart (Image credit: Marvel Studios/Disney+)

I already covered Muse's comic book history in my Daredevil: Born Again cast and character guide. But, here's another quick rundown on the masked individual with a penchant for the macaber.

In the comics, Muse is a nameless serial killer who uses his victims to create grotesque works of art. And, when I say "uses his victims", I mean that in the literal sense. Indeed, everything from their blood, skin, and other body parts are put to use by Muse to make stomach-churning paintings, sculptures, and other forms of art.

A recent addition to Daredevil's rogues gallery, Muse only made his comic book debut in Daredevil Volume 5 #11, which was released in September 2016. He was known to possess superhuman strength and agility, and had the ability to make himself imperceptible to Matt Murdock and his vigilante alias. That made Muse more difficult to track down.

Which actor is actually playing Muse? (Image credit: Marvel Studios/Disney+)

In Daredevil: Born Again, it doesn't seem like Muse will be superpowered. With the MCU favoring a more grounded, realistic take on Marvel's illustrious literary works, it appears Muse will be an ordinary (if that's the right word to use) serial killer who won't be able to go toe-to-toe with someone like Daredevil.

Indeed, Born Again's official trailer briefly showed Murdock's crimefighter making light work of Muse. Sure, it's only the tiniest of clips, so it's hard to determine if the MCU's Muse would stand any chance of holding his own in a duel with Daredevil. Based on that evidence, though, he won't.

Alright, so which actor is playing Muse? We still don't know. Four episodes in, Born Again's end credits sequence hasn't given us any indication who it might be. Marvel fans, myself included, have a wild theory about which Born Again character may be moonlighting as the murderous Muse, though. Hopefully, it won't be long before we find out if we're right.

Marvel Unlimited monthly subscription: was $9.99 per month now $4.99 at Marvel
The ultimate digital comics subscription for Marvel fans is offering new and returning customers 50% off their first month, which is much cheaper than a streaming subscription to watch all the best superhero movies. To gain access to a library of more than 30,000 comics, use the code 'SPIDEY50' at checkout. You're able to use the Marvel Unlimited app on all iOS and Android devices, including the web, too. But, be quick – the deal expires on May 5!View Deal

Who is Adam in Daredevil: Born Again and who portrays him?

Well, now we know what happened to Adam! (Image credit: Marvel Studios/Disney+)

This question is far easier to answer.

The Marvel Phase 5 TV show's first three episodes heavily implied that Wilson Fisk's estranged wife had an affair with a man named Adam. That happened while Wilson Fisk was recuperating after his surrogate daughter Maya Lopez shot him in Hawkeye, and then after she used her own superhuman abilities on Wilson to mess with his head in Echo's final episode.

'Sic Semper Systema' finally confirms that Vanessa had an extramarital affair with Adam while Wilson wasn't in New York City. It also reveals that Wilson Fisk tracked him down and, to paraphrase the city's newly-installed mayor, had a polite chat with Adam to prevent him from seeing Vanessa again.

Despite promising Vanessa that he'd never hurt Adam, though, the ever-deceptive Wilson Fisk hasn't exactly kept his word. In episode 4's final moments, it's revealed that Wilson Fisk has got Adam locked up in an undisclosed location. Not only that, it appears as though Fisk is denying Adam his basic human rights as he has no access to food, water, and clean clothes. Judging by Adam's appearance, he's been imprisoned for some time, too.

As for which actor is portraying Adam, that'll be Lou Taylor Pucci, who starred in 2013's Evil Dead remake, Apple TV+'s Physical TV series, and American Horror Story: 1984 on Hulu.

What does Born Again episode 4's title 'Sic Semper Systema' mean?

Episode 4's title has a deeper meaning for numerous Born Again characters (Image credit: Marvel Studios/Disney+)

Translated into English, the Latin phrase roughly means "thus always to the system".

There are multiple things we can read into its use in one of the best Disney+ shows' fourth chapter. The most obvious one relates to Matt Murdock, who's spent the past three-and-a-half episodes battling the US legal system to get justice for his clients.

Then there's Fisk, who continues to beat his head against a brick wall when it comes to the bureaucracy and red tape that stands between him and his mayoral pledges to get things done.

'Sic Semper Systema' also relates to Muse and Frank Castle/The Punisher, both of whom make their first MCU appearances in this episode. The pair are fighting the system in their own unique ways – Muse getting away with literal murder, while Jon Bernthal's anti-hero has railed against the status quo and doled out his own form of justice for years now. So, this chapter's title could refer to any number of characters in the show.

Who is Leory Bradford in Born Again? Episode 4's connection to Secret Invasion and The Marvels explained

No, Leroy Bradford isn't a Skrull masquerading as a human (Image credit: Marvel Studios/Disney+)

Amid everything else that goes on in 'Sic Semper Systema', the MCU weaves its interconnectivity magic once more by making a passing reference to two of the lowest-rated MCU projects of all-time.

At points throughout episode 4, Murdock meets Leroy Bradford, a petty thief who becomes the former's lates client. Bradford is facing another stint in prison after he's accused of stealing groceries from a local store.

In spite of Bradford's protestations that he didn't do it, Murdock confirms the evidence is firmly stacked against him to the overwhelming evidence that Bradford is the culprit. In a last-ditch attempt to prove his innocence, Bradford asks "could it have been a Skrull?", to which Murdock replies "nope".

Episode 4's reference to the Skrulls, aka the shapeshifting alien race we've seen in multiple Marvel projects, is pretty significant. It means their presence on Earth is no longer a secret. If a petty criminal like Bradford knows they exist, then all of humanity will.

Regardless, the Skrulls haven't exactly been used to their full potential in the MCU. The three projects they've appeared – Captain Marvel, Secret Invasion, and The Marvels – aren't exactly fan favorites. In fact, the latter two are considered to be among the worst MCU productions ever. You've got to wonder, then, if we'll ever see the Skrulls used in another Marvel movie or TV show.

When will the next episode of Daredevil: Born Again air on Disney+?

"Hey Matt, just to let you know that two new episodes will drop next week!" (Image credit: Marvel Studios/Disney+)

Episode 5 is set to be released on Tuesday, March 25 in the US. It'll air on Wednesday, March 26 in the UK and Australia.

It won't be the only one that does so, though. I previously reported on Daredevil: Born Again's unusual release schedule, which confirmed there'll be a two-parter next week. So, you'll be able to watch episodes 5 and 6 on Disney+, aka one of the world's best streaming services, on the aforementioned dates.

You might also like

• Here's how to watch the Marvel movies and TV shows in order
• Daredevil: Born Again episode 3 contains another Marvel reference to Spider-Man, but it's got nothing to do with Tom Holland's Peter Parker
• Get the lowdown on Marvel Phase 6

• Leaked European Samsung Galaxy S25 Edge prices suggest it will cost more than the Galaxy S25 Plus
• Storage configurations and colors have also leaked
• The predicted colors suggest the phone will have a titanium frame, like the Galaxy S25 Ultra

We’ve long predicted that the Samsung Galaxy S25 Edge could be positioned between the Samsung Galaxy S25 Plus and the Samsung Galaxy S25 Ultra, and now a pricing leak suggests we were right.

Android Headlines claims that – in Europe at least – the Galaxy S25 Edge will start at somewhere between €1,200 and €1,300, for which you’ll apparently get a model with 256GB of storage and 12GB of RAM.

We won’t bother with a price conversion here as those are rarely accurate, but for comparison, in France, you can get a Samsung Galaxy S25 Plus with that amount of storage for €1,172, while the Samsung Galaxy S25 Ultra starts at €1,472. So that Galaxy S25 Edge price would put the phone somewhere between the two, but closer to the S25 Plus’s price.

The Samsung Galaxy S25 Edge (Image credit: Lance Ulanoff / Future)

For reference, in other regions the 256GB Samsung Galaxy S25 Plus costs $999 / £999 / AU$1,699, while the Samsung Galaxy S25 Ultra starts at $1,299 / £1,249 / AU$2,149. So if the pricing above proves accurate then we'd expect to see the Galaxy S25 Edge cost somewhere between those two.

It’s a similar story for a version of the Galaxy S25 Edge with 12GB of RAM and 512GB of storage, which Android Headlines says will cost somewhere between €1,300 and €1,400.

That again would put it between the cost of the 512GB Galaxy S25 Plus and the 512GB Galaxy S25 Ultra, which elsewhere cost $1,199 / £1,099 / AU$1,899 and $1,419 / £1,349 / AU$2,349 respectively.

Three titanium shades

Android Headlines also shared the colors that the Galaxy S25 Edge will apparently be sold in, saying we can expect Titanium Icyblue, Titanium Silver, and Titanium Jetblack, which likely means titanium will be used in the construction of the phone.

We’d take all of this with a pinch of salt for now, but the Samsung Galaxy S25 Edge might be announced quite soon, with Android Headlines saying to expect it next month. Not all sources agree, but most leaks suggest we’ll see it by the end of May at the latest.

So hopefully we’ll know all the official specs and prices before too long, but in the meantime we’ll be sure to bring you all the credible leaks and rumors about Samsung's upcoming slim smartphone.

You might also like

• The Samsung Galaxy S25 Edge is being tipped to come with a sweet Google Gemini deal
• Best Samsung phones: top Galaxy handsets ranked
• Samsung Galaxy S25 Ultra review: the ultimate Android

A federal judge ruled that Elon Musk and DOGE likely violated the Constitution when shuttering the USAID. And, a look at when egg prices are likely to drop.

(Image credit: Chip Somodevilla)

• The Galaxy Z Fold 7 and Galaxy Z Flip 7 are due in July
• We might have to wait longer for the cheaper Galaxy Z Flip FE
• The phone could feature the Exynos 2500 chip

Samsung has got a host of foldable phones in the wings ready for later this year, but it's not clear exactly when we're going to be able to get our hands on them – and a fresh rumor suggests the Galaxy Z Flip FE could be the phone we're waiting the longest for.

According to South Korean outlet The Bell (via 9to5Google), the Samsung Galaxy Z Flip FE – believed to be a cheaper flip phone option – is going to launch several months after the more expensive Samsung Galaxy Z Flip 7, expected in July.

The main problem, as per the report, is that Samsung is struggling to make enough chips for the foldable. The company is said to be aiming to get its own Exynos 2500 chip inside the Galaxy Z Flip FE, but issues are continuing with production.

Those same production issues are reportedly part of the reason that the Samsung Galaxy S25 series went all-in with Qualcomm chips. The Exynos slices of silicon missed out with those flagship phones, and it looks as though they may miss out again with the Samsung Galaxy Z Fold 7 and the Galaxy Z Flip 7.

Display and specs

The cover display on the Galaxy Z Flip 6 (Image credit: Samsung)

While Samsung continues to toil away on Exynos 2500 production, it sounds as though the development of the Exynos 2600 is now well underway too (something that's been previously rumored). Samsung's aim is to make sure this next series is ready for the Samsung Galaxy S26 line-up.

In years past, Samsung has mixed up the chipsets used in its phones, depending on region and model: the Exynos chips are cheaper for Samsung to include, though they typically don't quite match the power of the Qualcomm Snapdragon alternatives.

As for the Samsung Galaxy Z Flip FE, we're anticipating that it's going to come with the same display as the Galaxy Z Flip 7, but a less powerful camera and overall specs. The phone has even been teased early by Samsung itself.

However, none of this is certain yet – Samsung's plans remain up in the air for its 2025 release schedule. Well-known tipster @MaxJmb suggests the Galaxy Z Flip FE may indeed launch alongside the Galaxy Z Fold 7 and Galaxy Z Flip 7, most probably in July.

You might also like

• Samsung may just have hinted at a Galaxy Z Fold 7 FE
• Galaxy Z Flip 7 leak gives us another key spec
• Our full review of the Samsung Galaxy S25 flagship

• Flashpoint report outlines a rise in stolen credentials and data breaches in 2024
• Over 3.2 billion credentials were compromised in 2024
• Profits from these attacks are fueling the cyber threat landscape

Over 3.2 billion credentials were compromised in 2024, new research has claimed, with these details then used to ‘fuel a number of illegal campaigns’.

Flashpoint’s Global Threat Intelligence Report said the figure represents a 33% spike year-on-year, and of these stolen credentials, an overwhelming 75% of these were obtained through information stealing malware, or ‘infostealers’ - which have now infected over 23 million devices worldwide. The most prolific of these is Redline - infecting 9.9 million hosts in 2024.

Unsurprisingly, 2024 saw a rise in ransomware attacks, with Flashpoint identifying a 10% increase across all sectors, adding to the 84% increase from the previous year - reinforcing reports that 2024 was the worst year for ransomware on record.

A resilient market

When ransomware and infostealer groups are disrupted, the market bounces back remarkably quickly, with new groups and malware strains emerging - indicating a “highly adaptable and resilient threat landscape”.

This makes life much more difficult for security teams and law enforcement, so robust threat monitoring and adaptability will be key for cybersecurity going forward.

A key takeaway from the report is the increase in data breach activity, with 6,670 publicly reported data breaches in 2024 - responsible for the exposure of over 16.8 billion records - putting organizations, customers, and users at risk, and funding cybercriminal organizations around the world.

“We live in interesting times, where technology is both a blessing and a curse,” said Ian Gray, VP of Intelligence at Flashpoint.

“The cybersecurity community will find new challenges in AI-augmented attacks, expanding attack surfaces, and increasingly sophisticated social engineering. Threat actors will find not just the opportunities they seek, but also the heightened scrutiny, improved defenses leveraging AI-tools shepherded by human intelligence, coordinated law enforcement actions, and inevitable consequences that follow.“

You might also like

• Take a look at our picks for the best malware removal software around
• Check out our choice for best antivirus software
• Free online file converters could infect your PC with malware, FBI warns

• A German promotional video for the Google Pixel 9a has leaked
• Numerous promo images have also been shared by a leaker
• Between them they reveal numerous likely specs and features of the phone

It’s hard to believe that there’s anything left to leak about the Google Pixel 9a, as in the last couple of weeks we’ve seen specs lists, pictures, and even videos emerge. But now, some seemingly official promotional materials have leaked, in both picture and video form.

First up, there’s a short promo video shared by reputable leaker Roland Quandt on Bluesky. The video is in German, but it provides a good look at the phone, complete with its widely rumored new camera design.

The video is particularly focused on showing off the Google Pixel 9a’s AI skills, including the presence of Gemini, and Google’s Add Me feature, which lets you combine multiple group photos so the photographer can also be in the shot.

9a pic.twitter.com/MKPoqNW3eEMarch 18, 2025

Next up, leaker @MysteryLupin has shared several promotional images showing the Pixel 9a and highlighting various specs and features. Those features include photography tools like Macro Focus mode and Best Take, the latter of which merges multiple group photos to get the best combination of facial expressions.

The images also mention other existing Pixel features, like various theft protection tools and Call Screen.

A long-lasting battery and drop protection

On the specs side, there’s apparently a 6.3-inch screen, a Tensor G4 chipset, 8GB of RAM, a 48MP main camera, a 13MP ultra-wide one, a 13MP front-facing camera, and the promise of over 30 hours of battery life. The images also mention water resistance and “enhanced drop protection.” We've largely heard this stuff before, but 30-hour battery life sounds promising.

Finally, the Pixel 9a apparently comes in Obsidian (black), Porcelain (white), Peony (pink), and Iris (blue) shades.

As ever we’d take these leaks with a pinch of salt, but not only do they come from reputable sources, they also look very convincing, so we expect these are genuine promotional materials.

We should find out very soon, as leaks suggest the Google Pixel 9a could land later today (March 19), so stay tuned for all the official details.

You might also like

• The Google Pixel 9a has gone up for sale and it’s not even out yet
• Multiple hands-on Google Pixel 9a videos have emerged, days ahead of the likely launch
• Best Pixel phones: these are the Google phones to buy right now

• The iPhone 17 Air will come with MagSafe, according to a new leak
• This is Apple’s magnetic charging tech that snaps your phone into place
• A leaked image of an iPhone 17 Air case has also surfaced

For months now, we’ve been hearing rumors that Apple will soon release an ultra-thin iPhone supposedly dubbed the iPhone 17 Air. Now, a new rumor has emerged regarding the device, and it’s good news for fans of wireless charging.

According to leaker Digital Chat Station on Chinese social media site Weibo, the iPhone 17 Air will support Apple’s MagSafe technology. This ensures optimal wireless charging efficiency by magnetically snapping the back of your iPhone to the charging puck.

Digital Chat Station used its post to back up rumors posted by leaker Sonny Dickson over the past weekend. Writing on Weibo, Digital Chat Station said that “The whole series has MagSafe,” referring to the upcoming iPhone 17 lineup, which would include the iPhone 17 Air should Apple unveil that phone.

That stands in contrast to the iPhone 16e, which came without MagSafe functionality when it launched. That was presumably to help keep the cost down, but with the iPhone 17 Air expected to be a mid-range phone when it arrives, that concern is presumably less pressing for Apple.

iPhone 17 Air case leaks out

(Image credit: TechRadar)

This isn't the only iPhone 17 leak we’ve recently come across. The aforementioned Sonny Dickson has posted a new photo of a case which they claim has been made specifically for the iPhone 17 Air. It features a clear back shell, and a large horizontal cutout for the camera and rear sensors.

Posting on X, Dickson said that “If you didn’t know an Air was coming, you’d swear it was a Google Pixel case.” Google’s Pixel phones have a horizontal bar along the back that houses the camera array, and multiple rumors have suggested that the iPhone 17 Air will adopt a similar look.

Apple is expected to unveil the iPhone 17 Air alongside the rest of the iPhone 17 range at an event in September. Until then, we can’t be sure exactly what that device will include; but if the latest rumors are true, both MagSafe and the horizontal camera bar could be ticked off your bingo card.

You might also like

• Apple expert just tipped a load of iPhone 17 upgrades: here are 7 things we’ve learned
• The iPhone 17 Air could have an affordable price, and better battery life than you might have expected
• The iPhone 16e doesn’t have MagSafe, but apparently Apple thinks you didn’t want it anyway

• Google has shared 6 health AI updates at its annual event The Check Up
• It is improving AI Search results for health queries, and helping researchers parse large volumes of literature
• It has also created a model that could improve AI-powered drug discovery

At its annual The Check Up event, Google has shared six ways it says it's using AI to improve health care and advances in medicine and science.

The company claims "AI can lead to scientific progress and cutting-edge products that help improve health outcomes for people all around the world." While some of the benefits of tools like Gemini are much more obvious when it comes to sifting through emails or doing research, the medical application can appear less obvious, although no less exciting.

Here are the six developments the company shared, including one that I think might cause some concern.

1. AI Overviews Search improvements

(Image credit: Google)

The first development is a change to Google's AI Overviews in Search, which I believe will have the biggest day-to-day impact on Google users and should be treated with the most caution.

Google says Search and AI Overviews "to find credible and relevant information about health, from common illnesses to rare conditions," and that it's improving the AI Overview results on health topics to be "so they’re more relevant, comprehensive and continue to meet a high bar for clinical factuality."

The change is a new What people suggest section. "While people come to Search to find reliable medical information from experts, they also value hearing from others who have similar experiences," Google says.

To that end, AI will organize different perspectives from online discussions to help you sift through helpful experiences from people in similar situations. The example Google uses is a person dealing with arthritis who might want to know how other people with the condition exercise.

Obviously, there's the potential for misinformation to surface here. Google's image includes a disclaimer that the results are "for informational purposes only" and suggests consulting a medical professional for advice or diagnosis.

As with everything you read on Google, a level of caution and discernment is required, all of this information already exists on the internet, Google is just trying to make the helpful stuff easier to find. Real-world results will determine whether or not it's successful.

2. Medical Records changes

Google has also launched a new Medical Records API globally in Health Connect, which lets apps read and write medical record information like allergies and medications in a standard format that you can share with your doctor's office.

3. Pixel Watch 3 Loss of Pulse Detection

Announced last month, the Pixel Watch 3, one of the best Android smartwatches, is getting Loss of Pulse Detection in the US at the end of March. The tool can automatically call emergency services and notify people close by if your heart stops beating.

4. AI co-scientist

Google's recently launched AI co-scientist can help researchers "parse large volumes of scientific literature and generate high-quality, novel hypotheses." Google says the tool won't automate the scientific process but is designed "to help experts uncover new ideas and accelerate their work." The company says it's already being used in Imperial College London and Stanford.

5. TxGemma

Google has launched a new collection of Gemma-based models it hopes "will help improve the efficiency of AI-powered drug discovery." The AI can "understand regular text and the structures of different therapeutic entities, like small molecules, chemicals, and proteins," which means researchers can use it to predict how safe or effective new therapies and drugs might be.

6. Cancer treatment

Finally, Google highlighted how it's helping a hospital in the Netherlands develop an AI tool that can "accelerate the identification of personalized cancer treatments by combining vast public medical data and de-identified patient data." It can reportedly generate "summaries of treatment options and relevant medical publications," giving Doctors more time to focus on patient care.

The efficacy and reach of all of these initiatives remain to be seen, but Google's update is a clear sign that as AI continues to permeate the world around us, its advance into every facet of life including medicine appears inevitable. With lives at stake and patient well-being on the line, getting it right is more important than ever, but the rewards for success are also infinitely greater.

You may also like

• Can ChatGPT really replace a therapist? We spoke to mental health experts to find out
• How AI-powered informatics and smart technologies could revolutionize healthcare
• The Apple AI health coach sounds like it's no longer stuck in development hell

• Ransomware operators Babuk adds Orange to their data leak site
• They claim to have broken into Orange in mid-March 2024, stealing sensitive data
• Researchers believe there is merit to the claims

Telecom giant Orange has reportedly again been hit by a ransomware attack after cybercriminal organization Babuk posted a data sample on its website, claiming proof of a successful breach at the company.

The group claims to have broken into Orange on Sunday, March 16, stealing “all information related to orange.com and orange.ro from Romania.”

“We will publish 1TB if they do not want to negotiate with us,” Babuk apparently said on its website. “And there is still a lot more that we stole, the sample is not much.”

"Very detailed information"

Babuk is not as popular as LockBit, or RansomHub, but it’s still a major ransomware player, who allegedly claimed 60 victims this year alone. It’s been around for years, although with long periods of inactivity.

If the group is telling the truth, they stole 4.5TB of “very detailed information”, including email addresses, customer records, source code, internal documents, invoices, contracts, projects, tickets, user data, employee data, messages, credit cards, call logs, and other personally identifiable information (PII) among the stolen data.

Researchers from Cybernews reviewed the sample posted on the website and say the claims “might be credible.”

“The threat actor uploaded a 6.44GB Orange data sample with thousands of Orange internal documents,” Cybernews explained. “Some files include employee data, like names, usernames, email addresses, and time zones, as well as a list of various Jira projects related to the Orange.ro domain.”

In late February, Orange Group confirmed suffering a cyberattack, but said at the time it was still looking into claims of valuable data being stolen. This attack was claimed by a member of the HellCat ransomware application, who also stole data belonging to Orange Romania.

Orange has not yet made any statement on the attack, but has been contacted for comment.

You might also like

• Telefónica says it was hit by systems breach, internal data leaked online
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

Changing from gas to climate-friendly electric appliances often involves expensive retrofits. A growing list of companies offer stoves, heat pumps and water heaters that make it easier and cheaper.

(Image credit: Jeff Brady)