News

Given the pressure that NHS doctors are under, it’s no surprise that many trusts are turning to AI for help. On the whole, this is a good thing.

AI tools can play a valuable role in managing workflows, automating documentation, coordinating care rotas and streamlining communications.

However, the pace of AI development, and generative AI in particular, has accelerated dramatically.

As a result, governance frameworks are failing to keep pace with frontline demand and healthcare practitioners have begun to use tools that have not been authorized by their organisation.

This is causing problems at all levels - particularly when those tools are used to support clinical decision-making.

“Shadow AI” is a term commonly used to describe the unauthorized use of AI within organizations. While the phenomenon exists across all industries, it has been growing rapidly in healthcare. One recent US survey revealed that 17% of healthcare workers admitted to using unauthorized AI tools in the workplace.

A another survey found that 39% of frontline healthcare staff say they use generic, free AI tools weekly or more. These are not isolated experiments. AI is already becoming embedded in day-to-day operations, often without formal governance or oversight, despite well-documented risks including hallucinations, inconsistencies and biases.

The reasons for this are understandable. Healthcare professionals are dealing with staff shortages, rising patient demand and overwhelming administrative workloads. Clinicians are expected to document more, communicate faster and process growing volumes of information - all while maintaining high standards of patient care.

It’s not surprising that they are reaching for the nearest AI tool. Particularly when that AI tool is extremely accessible, free or low-cost and easy to use. Shadow AI usage in the healthcare sector currently ranges from frontline clinicians right through to back office functions such as report drafting and recruitment support.

Most health professionals using “shadow” tools are not deliberately flouting rules, therefore. They are simply accessing some much-needed help from a convenient source. And their employer is not providing them with a robust alternative.

Medical organizations, and the NHS in particular, have often been cautious to roll out new technologies, though frequently with very good reason. Clinical systems require extensive testing, strong evidence bases and rigorous governance before deployment and decisions supported by AI can have direct consequences for patient outcomes.

Any clinical decision-making tool must therefore be thoroughly tested, based on solid medical research and exist under constant review and improvement. Yet delays allow shadow AI to fill the gap.

AI tools of the right quality are now reaching the market but the problem remains as to how leaders can rein in the use of shadow AI. It’s a problem that must be addressed quickly because the stakes are high.

The risks of unsanctioned AI usage include security breaches and privacy violations as well as the top concern - patient care.

In terms of data security and privacy, generic AI solutions are effectively “black boxes”. Ask ChatGPT whether sensitive health data will be safe in its hands, for example, and it will advise that it does not comply with health information regulations.

This is not surprising - patient information is “special category data” under GDPR and frequently requires additional protection - but it illustrates the danger of submitting too much information into an unauthorized app.

When it comes to patient care, generic AI solutions used to support clinical decision making carry an inherent risk: even when outputs appear credible, errors may be difficult to detect and assess. They may pull information from a broad range of sources which may be outdated, unaccredited or inaccurate.

A clinician in a hurry, referring to an unauthorized tool, is taking a big chance. If an unauthorized tool provides incorrect guidance relating to symptoms, medications, diagnoses or treatment pathways, the consequences could be severe.

The answer is not a blanket ban, which is likely to send unauthorized AI further into the shadows. The focus for healthcare leaders must be to establish clear guidelines on the use of AI and to invest in education and training regarding its risks and benefits.

At the same time, organizations need to provide clinically validated, secure and easy to use alternative tools. If approved systems are too slow to deploy or fail to meet operational needs, clinicians and administrators will inevitably continue to turn to unauthorized tools instead.

Governance structures are also an important part of the process. These should include IT teams, security specialists and operational management because AI deployment simultaneously affects patient care, compliance, workforce efficiency and organizational risk.

AI will undoubtedly become an increasingly important part of modern healthcare delivery. The organizations that succeed will not necessarily be those that adopt AI fastest, but which create trusted frameworks on which clinicians can access secure, medically robust and properly governed tools.

We feature the best Electronic Health Records software.

This article was produced as part of TechRadar Pro Perspectives, our channel to 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/pro/perspectives-how-to-submit

When I look for a gaming microphone, my requirements are modest – I want my teammates to hear me clearly and my foes to know what I really think of them (just kidding, those obscene thoughts stay in my head). I’d wager that most gamers have similarly unpretentious needs.

Razer, however, thinks gamers could do with more – a lot more. And so it’s brought out the Serien V3 Pro, a microphone that’s stuffed to the gills with high-end features that take it way beyond your common-or-garden gaming mic.

But to be fair, it’s not just positioned for gamers. Razer pitches the Seiren V3 Pro as the perfect partner for “musicians, podcasters, and creators who demand more than entry-level gear,” and it’s outfitted its “studio-quality” device with a range of features that might tickle the fancy of users a little more discerning than your average CoD goon.

That includes 32-bit float. Razer has included this expanded capture range to rescue audio that’s been subject to a few too many what we might call “gamer moments.” It can handle more extreme volume spikes – the type that might be prompted by jump scares, frustrating moments and clutch kills – with the idea being that streamers won’t have to worry about clipping and distortion during their broadcasts.

Razer has supplemented this feature with a fistful of other premium addons, including dual 48kHz and 96kHz sample rates, a built-in shock mount and removable pop filter, and a 30mm dynamic capsule. There’s a digital signal processing (DSP) that enables features like an audio expander, compressor, AI noise removal, and more. Razer has also included an XLR connector in addition to USB-C, just in case you want to run your mic through a dedicated audio interface.

Again, it’s all geared towards pros and enthusiasts rather than your average video game enjoyer.

Audio performance is obviously the most important part of any microphone – you can have all the bells and whistles, but if you end up sounding muffled or distorted, they’re not much use to you.

Here, the Seiren V3 Pro impresses. My audio was crisp and clear right out of the box, with no adjustments necessary. Razer walks you through a brief set-up process in Synapse when you get started, and it’s worth doing so the mic can be tweaked to your surroundings. The AI noise suppression in Synapse did a great job cutting out the whirr of my desk fan, while the pop filter put a stop to most unwanted plosives throughout my recordings and voice calls.

The on-device controls worked well, especially the mute button. It’s large and touch-sensitive, so it’s very easy to press in a flash (just be careful you don’t accidentally activate it). Both the button and the mic’s RGB ring turn red when you’re muted, giving a clear visual cue. Only a small nub of the gain dial protrudes from the underside of the device’s body, though, which can make it tricky to find.

As far as visuals go, the Seiren V3 Pro blends understated looks with a few choice touches. Its matte black finish and angled stand are easy on the eye, while this wouldn’t be a Razer product without a smattering of RGB lighting – in this instance, it’s about halfway up. Above it is the removable pop filter cover that hides the mic’s grille and built-in shock absorber. There’s a large touch-sensitive mute button on the front and ports for USB-C and XLR on the bottom alongside a 3.5mm headphone jack and a gain dial.

The build quality feels rock solid, with plenty of reassuringly weighty metal to be found. There’s a hefty foot to keep the mic grounded on your desk, but you can opt to screw the mic onto a boom arm if you prefer. The absurdly long USB-C cable (3.15m at our count) ensures you can trail it far from your PC.

That’s the hardware, but what about Razer’s notoriously finicky software? 32-bit float is only available through Razer’s app, so love it or loathe it, you’ve got to use it to get the full experience.

Synapse is full of controls to tweak the mic to your needs. You can adjust EQ filters and frequencies, change the vocal bass and ramp up the vocal exciter, mess with the stream mixer levels for input and output, enable a noise gate, reverb and AI noise suppression, and even dial in the RGB lighting if required.

To the average gamer, this is all a bit excessive unnecessary, even daunting. But if you’re a podcaster or musician – or just someone who wants as much control over their audio as possible – there’s plenty here to play with.

And that sums up the Seiren V3 Pro as a whole, really. The mic’s high-level features and equally lofty price tag mean it’s simply off the radar for most gamers. But if you know you need it and are willing to pay, you’ll find there’s a lot on offer here.

• Available to buy now
• Priced at $249.99 / £249.99 / AU$429.95
• Optional “Main Character Bundle” including the Seiren V3 Pro and Razer’s Kraken Kitty V3 Pro headset

The Razer Seiren V3 Pro is available now from Razer’s website or from third-party retailers. It’s priced at $249.99 / £249.99 / AU$429.95.

That makes it pretty expensive for a gaming microphone, so you should think carefully about whether you need all of its extra features before pulling the trigger. If you spend most of your time on video calls with family or yelling profanities at enemies in Call of Duty, you could probably make do with something more affordable.

Razer sells an optional “Main Character Bundle” that includes the Razer Seiren V3 Pro and the Razer Kraken Kitty V3 Pro headset. That’s sold for $429.98 / £429.98 / AU$759.90.

Polar patterns

Cardioid

Frequency range

50Hz – 16kHz

Connectivity

USB-C, XLR

Audio

24-bit / 32-bit (32-bit float via Synapse)

Sample rate

48kHz / 96kHz

Attribute

Notes

Score

Features

Lots of premium extras, plus plenty of control in the Synapse app

4.5

Performance

Excellent audio performance from the get-go

4.5

Design

Solid design, if unexceptional

3.5

Value

Expensive and much more than most people need, but decent value for pros

3.5

You’re concerned about loud noises on your recordings
With 32-bit float support, there’s more leeway to capture loud sounds without ruining your audioView Deal

You don’t like fiddling with settings
This mic sounds great right out of the box, making it ideal for anyone who wants to get up and running quickly. And there’s a guided setup process in Synapse in case you need itView Deal

You have an audio interface
With XLR connectivity, Razer’s Seiren V3 Pro can be hooked up to an audio interface for even more controlView Deal

You’re on a budget
There are definitely more straightforward options available for less cashView Deal

You want something simple
All those extra features are probably more than most gamers needView Deal

You need more than one polar pattern
The supplied cardioid polar pattern is great for streamers, but not so much for podcast hosts interviewing guestsView Deal

Blue Yeti

This classic mic is tried and true, with fantastic sound quality and superb ease of use. It sounds great out of the box and doesn’t require any additional apps to run. It’s also a fraction of the price of the Razer Seiren V3 Pro. That all makes it our pick for the best streaming mic money can buy.View Deal

SteelSeries Alias

The SteelSeries Alias is one of the best streaming microphones around. It comes with handy features designed for gamers, including a built-in shock mount and helpful mute indicator, plus an ingenious LED display that shows handy info (such as if your input volume is a touch too high). Like the Blue Yeti, it’s also more affordable than Razer’s offering.View Deal

• I created test recordings and also analyzed the mic’s performance during games and on audio calls with friends
• The microphone was tested on both my PC and my Mac
• I also spent time testing out the microphone’s software features in the Razer Synapse app

I spent a week testing the microphone on both my PC and my Mac. I created various audio recordings, used it during gaming sessions and put it through its voice-calling paces with friends.

• Read TechRadar’s reviews guarantee
• First reviewed: July 2026

Businesses are racing to adopt and operationalize AI, but many are deploying the technology faster than they can govern it. This gap is quietly becoming one of the biggest security risks facing enterprises today.

AI agents are being embedded into everyday workflows as teams chase efficiency gains, and organizations are under increasing pressure from boards and senior leadership to automate processes faster.

However, beneath this perceived momentum sits a more uncomfortable reality: businesses still have limited visibility into how these systems behave once deployed.

Security teams need to take a step back and rethink what proper governance looks like in a world where AI systems operate with autonomous access, inherited permissions and limited visibility.

Understandably, many businesses view successful AI deployment as evidence that they are ready to scale AI across their business.

Recent research found that 87% of organizations believe their identity management posture is prepared to support AI-driven automation. Yet, at the same time 46% admit that their identity governance falls short.

This contradiction sits at the core of the AI security paradox.

The core problem is that organizations are granting AI systems increasing levels of access and autonomy before they have established clear ways to monitor or verify how these systems behave.

If nobody can explain why an AI agent took an action, the permissions it inherited, the systems it accessed and whether or not that behavior was actually intended, governance quickly becomes reactive rather than preventative.

Unfortunately, securing these new AI identities won’t happen overnight, as legacy governance models simply aren’t cut out for the challenges facing security teams today.

Traditional identity management was built around four key assumptions: predictable behavior, human intent and bounded permissions. Agentic AI changes all four. Security teams already understand the risks of AI agents and non-human identities. The problem is that operational realities are forcing many organizations to accept those risks regardless.

Research shows that 73% of organizations believe standing access for AI agents increases security risk. In other words, leaders recognize the dangers associated with giving AI systems privileged access, but feel under pressure to prioritize speed and operational efficiency anyways.

Crucially, the governance issue extends beyond risk. The fact that 80% of organizations said they cannot always determine why an AI agent took a privileged action suggests a fundamental issue.

That creates an entirely different governance challenge from one that ever existed in traditional identity management. Humans can typically justify intent, context or decision-making -autonomous systems can’t.

AI usage in the workforce has long moved on from a select few employees experimenting with ChatGPT. Now, the technology is deeply woven into production systems, workflows and enterprise data.

The scale of the problem is significant - 53% of organizations say they regularly encounter unsanctioned AI tools or agents accessing company systems or data, but only 28% can detect shadow AI in real time.

In reality, businesses are effectively allowing unknown contractors to roam freely through their corridors, with limited insight into where they’re going, what they’re doing and what they have access to.

This shift is forcing organizations to rethink traditional security priorities and focus on continuous monitoring and validation.

As a result, businesses need to begin focusing on how effectively they can monitor and verify what their AI systems are accessing, and whether that access is still appropriate.

We know that organizations can’t afford to slow down on AI adoption given the competitive and operational incentives. But they also can’t afford to expand AI access and permissions faster than they can secure them.

This means bringing identity security to the forefront of AI governance strategies. Much of the discussion around AI risk still focuses on models, regulation and data privacy and those issues are important. But the true operational challenge increasingly sits between inherited permissions and unmanaged trust between systems.

For business leaders, the first priority should be improving visibility rather than adding more controls. They need to understand which AI identities exist within their organisation, what systems they can access, how permissions are inherited and where standing privileges have accumulated over time.

Organizations will also need to move away from persistent privilege models towards more dynamic models that provide temporary access only when required.

Ultimately, the organizations that will reap the greatest rewards from AI will be the ones that can understand, govern and validate the systems they use.

Deploying automation as fast as possible may lead to some quick wins, but doing so without the right governance measures in place could come back to haunt you.

We've reviewed and ranked the best firewall software.

This article was produced as part of TechRadar Pro Perspectives, our channel to 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/pro/perspectives-how-to-submit

With a price hike and frequent stock issues, it's been harder to justify picking up the Steam Deck in recent months. But if you're still on the hunt for a handheld gaming PC, then there are some solid alternatives more readily available — and at much lower prices. That includes this Lenovo Legion Go S at Amazon for a record-low price of £404.10 (was £549.99).

This exact same offer was available just a few months ago and it proved popular. With prices for PC components increasing and stock problems still lingering, I only expect that demand has increased since then.

That's why I think the Lenovo Legion Go S is a strong option right now. With an AMD Ryzen Z2 Go, 16GB of RAM, and a 1TB SSD, it's capable of running less demanding games, indies, and lots of modern favourites with ease. It also runs on SteamOS, which is designed specifically for the handheld form factor, making it easier and smoother to browse your library and play games compared to Windows 11 devices.

Put off by the price increase to the Steam Deck or tired of waiting for your desired model to come back in stock? I'd consider this Lenovo Legion Go S instead while it's back to a record-low price on Amazon. It offers solid power and performance thanks to the AMD Ryzen Z2 Go, 16GB of RAM, a 1TB SSD, and an eight-inch WUXGA display. You also get a device that runs on SteamOS, which is better designed for a smoother and more efficient handheld gaming experience compared to Windows 11.View Deal

I guess one of the only drawbacks to going for the Lenovo Legion Go S is the lack of an OLED display. Its eight-inch WUXGA display with a 120Hz refresh rate still offers smooth and vibrant visuals, though, all for around £250 less than the Steam Deck. I don't think it's worth paying that much more just for an OLED screen.

Yes, you may also have to make some compromises when playing the latest and most demanding games to get good performance, and battery life can be underwhelming, as with most handheld gaming PCs, but the Lenovo Legion Go S is a strong option right now as there are so few alternatives available at this price point.

We also rated it well in our Lenovo Legion Go S review, saying that it is "a compact beauty" and "a good choice for SteamOS diehards." Anyone with a large library of Steam games will find it a great device to play your favourite titles while on the sofa or on the go.

If you're upgrading a desktop or compatible laptop, this is a fantastic opportunity to add next-generation storage without maxing out your credit card.

Samsung's 9100 Pro 2TB SSD has crashed to $400 (was $680) at Amazon, knocking a massive $280 off one of the fastest consumer PCIe 5.0 drives around. For UK readers, the 9100 Pro is discounted to £364 (was £529) at Amazon, too.

Sequential read speeds reach up to 14,700MB/s, while write speeds climb to 13,400MB/s, delivering exceptional performance for demanding workloads. Everything you'll want to install and run will benefit from the extra bandwidth available through PCIe 5.0.

Offering up to 14,700MB/s read speeds, this PCIe 5.0 NVMe SSD pairs 2TB of storage with hardware encryption, advanced thermal control, Magician Software support, and excellent performance for creative workloads, AI applications, and demanding multitasking.

In the UK: now £366 (was £529)View Deal

In our review of the 9100 Pro we said: "Samsung's first true PCIe 5.0 drive is the best there is." In the summary we went on to add: "This is a very professional-specific drive in ways that previous Samsung Pro SSDs were not. If you're not a professional user, there are better PCIe 5.0 drives out there, but for pros, there's none better than the 9100 Pro."

Random performance reaches up to 1,850K read IOPS and 2,600K write IOPS, helping applications launch quickly and keeping multitasking smooth.

Samsung also focuses on efficiency, using a 5nm controller that improves power efficiency over the previous generation while advanced thermal controls help maintain consistent speeds.

Hardware encryption adds another layer of protection, and Samsung's Magician Software provides firmware updates, health monitoring, and extra management tools.

The M.2 2280 form factor fits a wide range of modern desktops and laptops that support PCIe 5.0 NVMe storage. With 2TB of capacity, there's plenty of room for creative projects, large software libraries without you needing to consistently manage free space.

Compatibility is straightforward for systems with an available M.2 PCIe 5.0 slot, and installation only takes a few minutes for most desktop builds.

If you've been waiting to benefit from PCIe 5.0 speeds, this deal delivers huge savings alongside outstanding performance, generous capacity, and useful software features that continue adding value long after installation.

The 9100 Pro SSD is also available in 1TB, 4TB and 8TB capacities at reduced prices.

The Samsung 990 Pro is one of the best SSDs around, especially at an unbeatable price like this. In our review, we hailed it as "an absolutely stellar M.2 SSD for both professional users and gamers."

In the UK: now £292 (was £503)View Deal

• US: Shop all internal SSDs at Amazon
• UK: See all internal SSDs at Amazon.co.uk

The Dreo Smart TurboPoly 765S is one versatile fan. It can be tall or short pedestal fan, or you can swap out the base and turn it into a desktop model. You can choose from a generous 11 speed modes, as well as six 'special modes', and horizontal or vertical oscillation (or both at once, turning it into a very effective air circulator).

You're spoiled for choice when it comes to control options, too — you can use the remote, touchscreen buttons on the top of the fan head, or adjust settings on your phone using the Dreo companion app.

In that app you'll find an absolute ton of customization options: I've tested some of the best fans on the market, and it's rare to see a fan that offers such precise control over when, where and how it blows. While Dreo does a good job of laying out the app so it's nice and useable, I do wonder how many people really need quite so many options.

So what about that basics? The Dreo Smart TurboPoly 765S delivers an even, effective flow of air, with a great range of settings. The lowest speed is whisper-quiet and gentle yet still perceptible, and the highest is impressively punchy. The Turbo setting — separated out as a special mode — is a miniature tornado. At higher settings, it's not the quietest fan I've used, but neither is it excessively loud.

You're paying a premium for the versatile form factor, as well as the extensive customization options and smart functionality. If you don't need all those things, there are cheaper, equally effective fans that'll fit the bill better. However, if flexibility and precise control is what you're after, the Dreo Smart TurboPoly 765S is a standout choice.

• List price: $159.99 / £169.99
• Available: US, UK

At list price, the Dreo Smart TurboPoly fan 765S costs $159.99 in the US and £169.99 in the UK (which equates to about AU$230, though at the time of writing it's not available in Australia). That puts in in the premium price bracket for fans, but it goes some way to justifying the cost with its versatile desktop-or-pedestal design and smart features.

That's especially true when you compare the wider market alternatives. For example, the original Shark FlexBreeze costs $199.99 / £199.99 / AU$249.99 at list price. The Shark option lacks smart features, but it can be used cordless (whereas the Dreo can't). Another alternative is the Meaco 1056P, at $189.99 / £149.99 (about AU$270). This is the model I have at home, and it's less versatile than the Dreo TurboPoly 765S, but quieter at higher speeds.

If you don't need app control and are happy to commit to either a desktop or a pedestal model, there are cheaper options available to you.

• Value for money score: 4 out of 5

Type:

Pedestal / tabletop

Speeds:

12

Oscillation:

150° horizontal, 100° vertical

Dimensions (W x H, floor mode):

11 x 39–44 inches / 27.9 x 99-111.8cm

Dimensions (W x H, desktop mode):

16.7 x 19.3 inches / 42.4 x 49cm

Control:

Buttons, remote, app

Cordless:

No

Special modes:

Auto, Turbo, Natural, Sleep, Custom

Max reach:

121.4ft / 37m

Timer:

Yes (12 hours)

• Versatile design can be a desktop or pedestal fan or air circulator
• App control is excellent; buttons / remote best for basic functions
• Wide speed range, precise oscillation control and multiple special modes

As fans go, the Dreo Smart TurboPoly 765S is one versatile performer. You can choose from a generous 11 speed modes, as well as six 'special modes', horizontal and vertical oscillation, and a timer option. You can direct the fan head at a specific target, or unleash the full range of oscillation to use it as an air circulator.

It can be used as pedestal fan, and in this mode it's possible to adjust the height to a taller or shorter floor-standing option. Alternatively, you can remove the base altogether and swap on a different one to turn it into a desktop fan. I especially like the design of the tabletop base — its opening mechanism protrudes so when you press it down on a flat surface the legs extend outwards to form a stable base.

The fan feels well-built and sturdy, and the pedestal base is reassuringly heavy, so I was never concerned about knocking it over. I also appreciated the inclusion of a moulded handle on the fan head, which makes it easy to carry it from the top.

Some competitors offer magnetic remotes that snap onto the fan itself when not in use, but that's not the case here. Instead there's a remote control holder that snaps onto the pole. It could feel more intentional but does the job perfectly fine.

There are a couple more advanced features the Smart TurboPoly 765S lacks. It can't be used cordless, so you'll have to commit to staying near a power outlet. There's also no misting capability — this is a newly popular concept in the fan world, and appears on a handful of modern fans including the Dreo 516S. It can be a bit of a game-changer if it's really hot.

That 'S' in the product code tells me there's an app I can use to control this fan. I can't overstate how jam-packed with options this app is. You can use it to turn the fan on and off, switch modes or adjust fan speed, and change the angle of the fan head. You can also precisely set the oscillation angles, including an option for asymmetrical oscillation.

There's a scheduling area where you can either choose when the fan will go on or turn off, and you can also set it to come on at a certain time (and speed / mode) on certain days. One minor missed opportunity is that you can't then schedule when it turns off — I wanted to set it to run for a couple of hours each evening as I go to sleep, but that doesn't seem to be possible.

The app is well laid-out and easy to use, although the fact there are so many precise customization options means it can feel fiddly at times. If you don't want to get your phone out, you can also control the fan using the remote control, or via touchscreen buttons on the top of the head.

These are responsive, and, to a point, straightforward. Changing the speed, swapping modes or adjusting the angle the head is pointing in is easy. However, for fine-tuning I found it far easier to use the app. (In fact, I had such trouble working out some of the functions that I emailed Dreo's general customer services address to help — and they get major brownie points for replying quickly, with a clear answer.)

• Design score: 4.5 out of 5

• A little noisy on higher settings
• Great range of settings, including an ultra powerful Turbo mode
• Easy to direct the air exactly where you want it

Overall, I was very impressed with the Dreo Smart TurboPoly Fan. The speed settings cover a good range, with level 1 delivering a gentle but perceptible flow of air, while level 11 is pleasingly powerful, and Turbo mode all-but blew me away. I can't imagine many situations where you'd need such strong airflow, but it's nice to have the option.

One small but welcome detail is the fact that the windspeed settings cycle from 11 straight to 1. That meant if I knew I wanted a higher setting, I could work backwards and save some button presses.

Also in its favor is how precisely you can adjust where the airflow is directed. In static mode, arrow controls enable you to easily change where the fan head is pointing. It's also possible to finely adjust the vertical and horizontal oscillation; even making it asymmetrical if you want. And of course, you can extend / retract the pole or swap from pedestal to desktop mode if you want to alter the height more dramatically.

I used it to keep me cool in my home office during two UK heatwaves, and adjusted it to a very narrow range of oscillation where it was essentially just blowing from side-to-side on my back. The effect was absolutely dreamy.

I also used it as an air circulator (in its full vertical and horizontal range of oscillation) to amplify the effects of our newly-purchased portable air con unit. It did an excellent job of rapidly distributing the cool air as it was being pumped in, filling the bedroom much quicker than the AC could manage on its own.

In terms of noisiness, I'd give the Dreo middling marks. On setting 1 it's whisper-quiet, but as I moved up the settings the sound ramped up, and on level 11 it was noticeably louder than my usual fan (a Meaco air circulator). It's certainly not the noisiest fan I've come across, though.

Moving on to the special modes. They are as follows:

• Normal: runs at fixed speed
• Auto: speed increases or decreases depending on room temperature
• Turbo: runs at maximum speed
• Natural: speed fluctuates to mimic a natural breeze
• Sleep: decreases speed after 30 mins, up to two times
• Custom (via Dreo app): speed changes at specified times or temperatures

I was pleased to see there's a Sleep mode included, and the setup for this one is pretty standard, allowing you an hour to drop off. It's worth pointing out that the lights on the Dreo Smart TurboPoly 765S switch off automatically as standard a few seconds after you've finished adjusting them — something I appreciated as I hate to have a glowing LED illuminating my room as I'm trying to sleep. (You can change this setting if you prefer the display to stay on.)

Auto mode is useful, too, if you're using the fan throughout the day. The ambient temperature measurement seems accurate, and matched that of my other fan.

I wasn't so enamored with the Breeze mode. I've seen this mode listed on a few different fan models recently, and have always been curious about it. On test, though, I found the fluctuations in speed and accompanying shifts in noise distracting more than anything. That said, it did feel somewhat like a natural breeze, and I can imagine some people enjoying this effect.

• Performance score: 4.5 out of 5

Attribute

Notes

Rating

Value

Premium pricing, somewhat justified by its versatility and smart functionality.

4 / 5

Design

Versatile design — can be used at a range of heights, controlled in a number of ways, and with plenty of setting options to explore.

4.5 / 5

Performance

Great range of airflow settings, including an ultra-powerful Turbo mode. A little noisy on higher settings, and controls can feel fiddly.

4.5 / 5

You need a tabletop and a pedestal fan

The Dreo Smart TurboPoly 765S can be used as a pedestal fan with a telescopic pole, or the base can be switched to turn it into a tabletop model.

You want lots of precise control

This fan is not lacking in customization options — there are plenty of ways to change exactly how, when, and where the Dreo TurboPoly 765S unleashes its airflow.

You'd like to be able to control the fan remotely

Smart functionality means you can operate this fan via a companion app on your phone.

You just want a tabletop OR pedestal fan

There are cheaper, as effective options if you don't need the 2-in-1 functionality.

You like to keep things simple

Some people won't want all the various control options, in which case there are cheaper and more straightforward fans that would fit the bill better.

I tested the Dreo Smart TurboPoly 765S over three weeks in June/July, which happened to encompass two UK heatwaves, with temperatures reaching 37C or higher (around 98F). I used it extensively during the day while working from home, as well as at night, and tested all the special modes. I also tried out all the control methods, including digging into the app. I compared my findings with other fans I've used, but predominantly my trusty Meaco air circulator.

Read more about how we test

• First reviewed July 2026

• Exposed ElasticSearch cluster at Nextcloud contained ~367k records (8GB), including employee data, client contracts, and scripts
• Sensitive information such as staff emails and client company details was left unencrypted; Nextcloud secured the archive within two days after notification
• Company attributed the incident to hosting misconfiguration, stressing customer servers were unaffected, though researchers warn attackers may have accessed the data

European cloud provider Nextcloud kept an unprotected database on the public internet, exposing sensitive internal and client data to anyone who knew where to look, experts have revealed.

Nextcloud is a free, open source platform that lets users create their own private cloud. It is often described as an alternative to Google Drive, or Microsoft 365, which allows users to control where their data sits.

In mid-May 2026, security researchers from Cybernews discovered a publicly exposed ElasticSearch cluster and, after a deeper investigation, determined it contained around 367,000 records (8GB of data in total). The archive was a mix of Nextcloud employee data, client company data, contracts, and scripts built for the company’s clients.

The majority of files were in .PDF format (71k), followed by .PNG (53k) and .MD (23k). All of the exposed records were found in a single index, with some revealing client company information, as well as data on Nextcloud staff. Some of the information was unencrypted, as well, exposing employee email addresses, client company names and addresses, and emails of individuals who sent invoices to Nextcloud.

Cybernews reached out to Nextcloud and the company locked the archive down within two days, and notified relevant authorities. It says it found no evidence of unauthorized access, but without a deep forensic analysis, it is impossible to say if that really is the case.

“If our team managed to discover the exposed dataset, threat actors may have too,” the Cybernews team wrote. “Malicious attackers operate numerous bots on the web that scour the net looking for exactly that: misconfigured databases with data to steal.”

The company also said this was a misconfiguration issue and that its services are secure: “The issue was caused by a misconfiguration of our hosting infrastructure and is not related to the Nextcloud solution. No other Nextcloud servers belonging to our customers, partners or other users have been affected by this issue,” the company’s spokesperson told the researchers.

The deadline for high-risk AI compliance under the EU AI Act may well have been pushed back. And for overstretched compliance teams, that might sound like welcome relief. But the organizations that will thrive in the age of AI are the ones that understand that an arbitrary deadline was never really the point.

The race to deploy AI is already in full swing.

Models like Claude Mythos and OpenAI’s GPT-5.5, dubbed "Spud", represent extraordinary leaps in capability. In the right hands, they accelerate decisions, surface insights, and create genuine competitive advantage. In the wrong hands, or without proper oversight of the data feeding them, the consequences can be catastrophic.

As Pocket OS discovered recently when an AI agent wiped out the company’s entire database in nine seconds.

When you strip away the regulatory language, what the EU AI Act is really asking for is provability. Article 10 sets out that high-risk AI systems, those used in credit scoring, insurance underwriting, hiring decisions, and similar consequential applications, must be built on training data that is traceable, well-governed, and demonstrably free from bias.

Organizations must document data origins, every transformation applied, the assumptions made, and how potential biases were identified and addressed.

This is not a box-ticking exercise. It is a fundamental rethinking of how organizations manage their data estates. And the penalties for getting it wrong are steep: up to €35 million or 7% of global annual revenue.

The underlying problem is that most enterprise data infrastructure was not designed with this level of traceability in mind. GDPR established guardrails around data storage and access, but the AI Act raises the bar considerably.

You now need to trace data from its original source, through every transformation, to its final impact on model outputs. Anecdotally, AI model validation can take between nine and 12 months, and that assumes you have the lineage infrastructure in place to begin with.

Nowhere is this more consequential than in financial services. Credit scoring models trained on historical data can encode the biases of the past, automating financial discrimination at scale, often without any individual in the organization realizing it is happening. A bank that cannot trace how a training dataset shaped a model’s outputs is not just exposed to regulatory risk. It is potentially perpetuating systemic harm.

The irony is that the sector has been here before. BCBS 239 already requires financial institutions to demonstrate data accuracy, integrity, and the ability to aggregate risk data on demand. The AI Act is not a new category of challenge, it is an intensification of one that rigorous data lineage was already built to address.

The organizations best positioned for this moment are those that have stopped thinking about data lineage as a compliance cost and started treating it as core infrastructure. There is a crucial distinction between these two perspectives.

The focus must now be a governance model that catches problems during design, before a model ever reaches production, rather than during an incident response. Lineage makes this kind of proactive governance practical at enterprise scale.

Bi-temporal lineage allows teams to recreate the exact data state used for model training at any point in time, which is essential for audit purposes. More importantly, it allows teams to simulate the downstream impact of a data or schema change before it happens, preventing the kind of silent model degradation that can quietly undermine ROI.

The deadline extension offers time, but not as much as it might appear. US regulators are already integrating AI into supervisory examinations. Gartner predicts that by 2028, 50% of organizations will adopt zero-trust data governance as AI-generated content proliferates through enterprise data supply chains. The regulatory direction of travel is clear and consistent.

Getting AI governance right requires building the lineage layer first. That is not a project that happens in weeks. And the organizations that start building now, not because a deadline forces them to, but because they understand what is at stake, will be the ones deploying AI at scale with confidence, auditability, and genuine trust.

The EU AI Act just got a stay of execution but organizations that treat it as breathing room may be sleepwalking into a far bigger problem.

We've featured the best AI tool.

This article was produced as part of TechRadar Pro Perspectives, our channel to 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/pro/perspectives-how-to-submit

• Sony has announced that it plans to end the production of physical game discs in January 2028
• The move has angered gamers, with a petition urging the company to reverse the decision racking up more than 250,000 signatures
• Major UK second-hand retailer musicMagpie warns that it could lead to higher prices and less choice for shoppers

Last week, PlayStation maker Sony announced that it would end the production of new game discs in January 2028 in response to demand for digital releases "significantly" outpacing that of physical ones. The move immediately angered gamers, with some vowing to stop buying PlayStation products entirely and creating a petition urging the company to change its mind.

The plea has now racked up almost 250,000 signatures, prompting massive UK second-hand retailer musicMagpie to weigh in.

Its in-house physical game expert, Alex Black, warns that Sony's decision will lead to higher prices and less choice for shoppers: “If the industry moves away from physical media entirely, consumers will have fewer opportunities to shop around for the best price or recoup some of the cost of new releases by trading games in."

He thinks that the decision even risks making the hobby less accessible by pricing out those with lower disposable income. "Pre-owned games give people access to more affordable gaming while extending the lifespan of physical products," he explains.

"Customers regularly buy pre-owned games to save money and sell completed titles to help fund their next purchase. It's a model that benefits both consumers and sustainability by keeping perfectly good products in circulation for longer."

The retailer is also aware of the online backlash to the announcement, and reckons that it shows how important physical releases are for gamers — even if Sony seems to think otherwise.

“The comments from gamers about Sony’s decision to go exclusively digital clearly demonstrates that there is a huge community of players who still value physical ownership, whether that's for collecting, preserving, or simply owning something they can keep, lend, or resell," Black says.

“For many consumers, buying a physical game means purchasing an asset that retains, or even increases, in value. Comparatively, digital purchases are tied to licences and various platforms, meaning consumers have less flexibility over what they buy."

Sales data from musicMagpie certainly indicates that there is still some demand for physical titles. Over the last year, the retailer sold roughly 360,000 physical video games, more than 157,000 of which were for PlayStation systems.

Although musicMagpie taking a stand might win points with gamers, it could already be too late for Sony to change course. The company has already made moves to repurpose the historic PlayStation factory that manufactured 24 billion discs, while industry analysts say that digital sales are "just too lucrative" to ignore.