Feed aggregator

• Microsoft created an AI red team back in 2018 as it foresaw the rise of AI
• A red team represents the enemy; and adopts the adversarial persona.
• Latest whitepaper from the team hopes to address common vulnerabilities in AI systems and LLMs

Over the past seven years, Microsoft has been addressing the risks in artificial intelligence systems through its dedicated AI 'red team'.

Established to foresee and counter the growing challenges posed by advanced AI systems, this team adopts the role of threat actors, ultimately aiming to identify vulnerabilities before they can be exploited in the real world.

Now, after years of work, Microsoft has released a whitepaper from the team, showcasing some of its most important findings from its work.

Microsoft’s red team whitepaper findings

Over the years, the focus of Microsoft's red teaming has expanded beyond traditional vulnerabilities to tackle novel risks unique to AI, working across Microsoft’s own Copilot as well as open-source AI models.

The whitepaper emphasizes the importance of combining human expertise with automation to detect and mitigate risks effectively.

One major lesson learned is the integration of generative AI into modern applications has not only expanded the cyberattack surface, but also brought unique challenges.

Techniques such as prompt injections exploit models' inability to differentiate between system-level instructions and user inputs, enabling attackers to manipulate outcomes.

Meanwhile, traditional risks, such as outdated software dependencies or improper security engineering, remain significant, and Microsoft deem human expertise indispensable in countering them.

The team found an effective understanding of the risks surrounding automation often requires subject matter experts who can evaluate content in specialized fields such as medicine or cybersecurity.

Furthermore, it highlighted cultural competence and emotional intelligence as vital cybersecurity skills.

Microsoft also stressed the need for continuous testing, updated practices, and “break-fix” cycles, a process of identifying vulnerabilities and implementing fixes on top of additional testing.

You might also like

• These are the best AI phones with cutting-edge tech
• Take a look at our guide to the best AI website builders
• AI is becoming increasingly vital in software development

• The US Treasury has revealed its foreign investments office was hit in recent cyberattack
• The office reviews national security risks for investments
• Breach of a third-party vendor led to US Treasury systems being accessed

The US department which reviews foreign investments for national security risks has been revealed as a further victim of the cyberattack which targeted the US Treasury department earlier this month.

The attack was declared a ‘major incident’ after a third-party cybersecurity service provider was compromised, allowing a threat actor remote access to key Treasury systems.

The Committee on Foreign Investment in the US (CFIUS) suffered a data breach as part of the campaign, CNN has revealed. The department reviews foreign investments for national security risks, and was recently granted the authority to scrutinize real estate sales near US military bases, with the scope to block Chinese investment into the US.

Carefully chosen targets

The news is the latest in a string of developments following the US Treasury department hack, in which hackers were able to gain access used by the breached vendor to override parts of the Treasury department’s systems.

The attack has sparked serious concerns for US officials, who have reportedly grown increasingly worried that the Chinese government or proxies plans to use land acquisitions to spy on US bases.

In the broader context of the Treasury attack, other targets look to have been chosen with China-US relations in mind. For example, the US sanctions office was targeted - which last week issued a sanction to a Chinese firm for its alleged role in cyberattacks.

The cyber espionage campaigns launched against US and western targets in recent months look not just to steal information and access sensitive data, but also to disrupt critical infrastructure.

In a separate recent attack, Chinese group Salt Typhoon allegedly breached 9 major telecommunications firms in a huge campaign against US critical infrastructure. Amongst the victims were Verizon, AT&T, and Lumen Technologies, who had threat actors lurking inside their networks for months.

You might also like

• Take a look at our pick of the best malware removal software
• Japan says Chinese hackers have launched hundreds of attacks against targets in the country
• Check out our pick for best antivirus software

The game, first announced in 2020, is available for gamers to try out on PC and on consoles.

• Hackers recently claimed to having breached Telefonica and stealing sensitive information
• The company confirmed the breach and said it was investigating
• The crooks broke into an internal ticketing system

Spanish telecommunications giant Telefónica has confirmed suffering a cyberattack, and said it was currently investigating its scope and potential damage.

The news comes after cybercriminals posted a new thread on a dark web forum containing a Jira database allegedly belonging to the company. One of the attackers spoke to BleepingComputer and said that the “internal ticketing system” is an internal Jira development and ticketing server which Telefónica uses to report and resolve internal issues.

The miscreants used compromised employee accounts to steal 2.3GB of documents, tickets, and other information.

The Hellcat connection

Apparently, the company was breached by four cybercriminals, three of which form the Hellcat Ransomware operation. This group emerged in late 2024, and quickly made a name for itself after targeting Schneider Electric and stealing more than 40GB of information.

In a peculiar move, Hellcat demanded a ransom of $125,000 to be paid in "baguettes," a term they later clarified as referring to Monero cryptocurrency, likely using the term "baguettes" to gain media attention.

Telefónica has now confirmed the breach, telling BleepingComputer it was currently looking into the matter.

"We have become aware of an unauthorized access to an internal ticketing system which we use at Telefónica," the company told the publication. "We are currently investigating the extent of the incident and have taken the necessary steps to block any unauthorized access to the system."

Telefónica is a leading Spanish telecommunications company with a global presence, serving approximately 392 million customers across Europe and Latin America. It operates in countries including Spain, Germany, the United Kingdom, Brazil, and several others in Latin America. In the first half of 2024, Telefónica reported revenues of $22.03 billion, with a net income of $1.06 billion, reflecting a 28.9% increase compared to the same period in the previous year.

The company maintains a strong market presence, with mobile market shares ranging from 18.2% in Mexico to nearly 54% in Venezuela as of 2023.

Via BleepingComputer

You might also like

• Schneider Electric says developer platform was breached, company data stolen
• Here's a list of the best antivirus tools on offer
• These are the best endpoint protection tools right now

Damaging winds gusting between 55 and 70 mph are forecast, increasing fire risks across LA and Ventura counties over the next few days. More than 150,000 people are already under evacuation orders.

(Image credit: David Pashaee)

• Samsung has just announced two new Galaxy Ring sizes
• It has also expanded the Ring to 15 new countries
• The rollout is accompanied by a sleep upgrade to the Samsung Health app

Samsung has today announced that the Samsung Galaxy Ring is now available in two new sizes and is rolling out to 15 new markets in February.

The best smart ring on the market, the Samsung Galaxy Ring now comes in US ring sizes 14 and 15. Along with the additional sizes, Samsung is rolling out the Galaxy Ring to 15 new markets, bringing the total to 53. Starting in February, the new markets include: Cyprus, Czech, Greece, Hungary, Israel, Japan, Malaysia, Mauritius, New Zealand, Singapore, Slovakia, South Africa, Taiwan, Vietnam, and Zambia.

Prospective buyers and existing owners alike will also benefit from an upgrade to the Samsung Health app focused on sleep and health goals for daily life, here's what's new.

Samsung upgrades its health app

(Image credit: Zachariah Kelly / TechRadar)

Samsung says its new health app can analyze your sleep environment, taking data from SmartThings homeware to assess factors like temperature, humidity, air quality, and even the light intensity in your room. You'll get a Sleep environment report to help optimize those conditions and use SmartThings to automatically adjust your surroundings and help you get a better night's sleep.

There's also new Sleep time guidance that will suggest optimal bedtimes tailored to your unique sleeping patterns, habits, and conditions, as well as provide actionable insights for better sleep.

Finally, there's a new Mindfulness tracker to help you improve your mental health, partly in aid of getting a better night's sleep. You can monitor your moods and follow breathing exercises and meditation in the app.

Given the Samsung Galaxy Ring is one of the best sleep trackers out there, these updates make it an even more potent bedtime companion.

You may also like

• I spent 24 hours with the Samsung Galaxy Ring, and it's scarily good so far
• The Samsung Galaxy Ring is brilliant, but it misses the mark in three key areas
• Why the Samsung Galaxy Ring, 2024's biggest (and smallest) wearable, is our Health & Fitness Device of the Year

In my grandfather's jewelry workshop, I witnessed the toll of repetitive tasks on human creativity—a challenge that inspired me to reimagine workplace dynamics through humanoid robotics that complement rather than compete with human workers.

The debate around robots in the workplace often gravitates toward a false dichotomy: human versus machine. As the founder of Humanoid, the U.K.'s first AI humanoid robotics company, I've observed that this narrative misses the transformative potential of human-robot collaboration.

Human-machine collaboration

In reality, we’ve already seen a number of successful examples of human-machine collaboration across a number of industries: manufacturing, automotive sector in particular, healthcare, and hospitality.

We’re seeing collaborative robots increasingly enhancing efficiency and customer service experiences across these industries, from Kime robots serving drinks at self-contained kiosks in Spain to BMW’s Spartanburg plan using cobots for door assembly tasks.

Another example is ANYmal from Zurich-based Robotic Systems Lab (RSL). Their product ANYmal is a quadrupedal robot designed for autonomous operation in challenging environments like industrial inspection. It can navigate complex terrain and carry specialized sensory equipment while weighing less than 30kg.

RSL has also developed SpaceBok, a lightweight quadrupedal robot designed to research dynamic legged locomotion for space exploration, featuring modular components for various experiments and optimized for explosive jumping maneuvers. In the field of environmental monitoring, RSL's ANYmal Dodo robot recently completed a challenging hike, demonstrating the lab's advancements in creating robots capable of traversing difficult outdoor terrains.

Across industries, the promise and potential of human-robot collaboration includes not just increased efficiency and productivity, but also improved safety, more informed decision-making, greater flexibility, 24/7 operations, reduced costs as well as great ability to multitask and for humans to focus on higher-value priorities.

For example, a Dutch startup Monumental, has raised $25 million to develop AI-powered autonomous robots for heavy-duty construction tasks like bricklaying. The idea is that their robots will help address labor shortages and improve efficiencies by working alongside human workers to automate mundane, repetitive tasks.

The European approach

The European approach to robotics stands apart from international counterparts through its emphasis on barrier-free integration and skills preservation. While other global players may focus on automation for efficiency gains, European initiatives like the groundbreaking Sharework project and SHERLOCK showcase a distinct philosophy centered on human-robot cooperation, knowledge transfer from aging workers, and democratized access to automation technology.

According to a recent MIT study, aging populations are a major driver of robotics adoption in workplaces, with demographic aging accounting for 35% of the variation in robot use among countries as companies implement automation primarily to fill labor shortages caused by an aging workforce rather than purely for technological advancement. This trend, particularly in Europe, has evolved toward collaborative human-robot workplaces where robots are designed to preserve and amplify human expertise rather than replace workers, enabling aging workers to transfer their knowledge while robots handle physically demanding tasks.

This approach prioritizes preserving craftsmanship while making manufacturing more inclusive and accessible to small and medium enterprises.

Take, for instance, the Sharework project, where robots work alongside humans without protective barriers, as demonstrated by Goizper Group's implementation in the Basque Country, where collaborative robots were successfully integrated into their servo rotary table assembly area for power transmission components.

This isn't just a technological achievement: It's a fundamental shift in how we perceive workplace safety and collaboration.

Cobots

These so-called cobots or collaborative robots are equipped with advanced sensors and AI systems that allow them to understand and respond to human movements, making them safe enough to share the same workspace as their human colleagues.

The psychological impact of this barrier-free collaboration is profound. When workers see robots as tools that enhance their capabilities rather than threats to their employment, workplace anxiety decreases and job satisfaction increases.

At Humanoid, we believe that employees working alongside humanoid robots will be more empowered to focus on creative and strategic tasks, knowing that repetitive or physically demanding work is handled by their robotic colleagues.

The SHERLOCK initiative, another European initiative, takes this further by developing robots that can learn from human expertise. Instead of programming robots with rigid instructions, these systems observe human workers, learning their techniques and adapting to different scenarios. This approach preserves and amplifies human craftsmanship rather than replacing it—a crucial distinction in industries where skill and precision matter.

Democratizing manufacturing

What's particularly exciting is how this collaborative approach is democratizing manufacturing.

Traditional industrial robots required extensive programming expertise and significant infrastructure changes, effectively limiting automation to large corporations.

Modern humanoid robots, designed for human collaboration, are more adaptable and user-friendly. This means small and medium-sized enterprises can now access automation technologies.

Moreover, this collaborative approach is opening up manufacturing to a more diverse workforce. By having robots handle physically demanding tasks, we're removing traditional barriers that may have prevented certain individuals from entering manufacturing roles.

The European model of human-robot collaboration is also addressing a critical challenge: the preservation of manufacturing expertise. As experienced workers retire, their knowledge can be captured and transferred through robots, ensuring that craftsmanship and industry-specific skills aren't lost to time. These robots become not just tools but repositories of human expertise, working alongside newer employees as both assistants and training aids.

The future of work

Looking ahead, the future of work isn't about choosing between humans or robots—it's about creating synergies between both. As we continue to develop more sophisticated humanoid robots, our focus remains on enhancing human potential rather than replacing it. The workplace of tomorrow will be one where humans and robots collaborate seamlessly, each bringing their unique strengths to the table.

This isn't just automation—it's an augmentation of human capability. And in the future, the question won't be whether robots will take our jobs, but rather how they can help us do our jobs better, more safely, and with greater satisfaction.

We've featured the best small business software.

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

Microsoft has a grand plan to revamp Edge apparently, and one of the possible next steps in the evolution of the web browser has just been flagged up.

As Windows Latest explains, the software giant has a project with the codename of ‘Phoenix’ which has the aim of overhauling Edge, and making it look more modern, but not in one big update – rather a series of drip-fed moves.

And one of the incoming changes could be the ability for the browser to morph its color palette to match Windows 11’s settings, the tech site theorizes.

This is based on a couple of flags Windows Latest noticed in a test build of Edge from the Canary channel (the earliest preview builds).

The flags are ‘msEdgePhoenixColorPaletteGenerator’ and ‘msUseSnappingPointsForPhoenixPaletteGeneration’, so we have the clear reference to the Phoenix project here.

What these flags likely refer to is the ability for Edge to run a piece of ‘color palette generator’ code that could detect Windows 11’s accent colors or theme, and change the browser to match that. In other words, Edge would dynamically switch to complement the Windows 11 environment in this respect.

(Image credit: Shutterstock / monticello) Analysis: What happens in testing, sometimes stays in testing

This is educated guesswork on Windows Latest’s part, but it makes sense to me. Bear in mind that not only is this early testing, but the functionality isn’t even enabled in preview yet, and only present in these flags hidden away in the background. It’s possible that nothing may come of this ultimately, but we can keep a watchful eye out for further clues.

Mica is also apparently coming to the Edge title bar, which is an opaque effect that means the bar is slightly tinted by the desktop color behind it.

While all this is commendable – if it happens – will it be enough to propel Edge forward in terms of browser share? I’m looking for bigger moves than this, and expect you will be too, although it certainly won’t hurt the chances of Microsoft’s browser.

To be fair to the software giant, some nifty bits have been brought in as part of Project Phoenix in the past, such as Edge’s split-screen feature which lets you browse two websites simultaneously, side-by-side in the browser window. Now that is a smart touch.

A further point to note here – albeit more minor – is another flag Windows Latest spotted recently in testing, namely ‘msVisualRejuvTooltips’ which likely represents a revamp of Edge tooltips. While this has since been removed, it’s still something Microsoft could return to at some point down the line.

Furthermore, Leopeva64, a source of a lot of Edge rumors and leaks on X, recently flagged up that the ‘Auto Picture in Picture’ feature in Microsoft’s browser is being removed from testing (see the below post).

Edge's "Auto PiP" feature still works in Beta and Stable (it no longer works in Dev and Canary), if you want to try it out before it stops working,here's the CLF to enable it: --enable-features=msEdgeAutoPictureInPicture,msEdgeAutoPictureInPictureTriggerhttps://t.co/dqiIlcXTq3 pic.twitter.com/UDu8LUPY7dJanuary 12, 2025

All of which illustrates exactly why changes in testing need to be regarded carefully – not all of them make it through to release by any means.

Whatever happens, it’s clear Microsoft is hoping to improve Edge substantially with Project Phoenix, but I’m hoping to see some bigger pieces of that puzzle come into view this year.

You might also like

• Microsoft plans to debloat Edge browser and take a leaf out of Google Chrome’s book when it comes to Settings
• Microsoft Edge is better than Google Chrome, but I can't bring myself to switch
• How to use Microsoft Edge

• Severance's creator has debunked one wild fan theory ahead of season 2's debut
• Dan Erickson humorously labeled the particular hypothesis as "ridiculous"
• He also suggested that the hit Apple TV Original won't answer every question that fans have

Three years have passed since Severance's first season ended, so fans have had plenty of time to come up with some outlandish theories about the hit Apple TV Original.

Of the seemingly infinite amount of online fan theories that have emerged, though, there's only one that creator Dan Erickson has amusingly called "ridiculous". Indeed, speaking to me ahead of Severance season 2's debut this Friday (January 17), Erickson debunked one wild fan hypothesis that, surprisingly, actually has ties to the man himself.

Asked for some examples of the strangest comments or theories he's seen or heard, the Apple TV Plus show's head writer replied: "Well, I've heard people say that they think I based Rick [played by Michael Chernus] on myself, which I think is ridiculous! Hey, just because he's well-dressed writer, it doesn't mean I've written myself into my own show."

For those who need a reminder: Rick – full name Ricken Hale – is the husband of Devon, aka Mark Scout's sister, and a published self-help author. His in-universe book, 'The You You Are, A Spiritual Biography of You', becomes something of a manifesto to Lumon Industries' Macrodafa Refinement team, of which Mark's 'innie' is a part of, who – spoilers! – eventually and openly revolt against the clandestine biotechnological megacorporation in the season 1 finale. Despite what some fans believe, though, Erickson didn't serve as the inspiration for Rick, so that's one conspiracy theory that we can strike off the list!

'Part of the fun is letting people speculate and come up with their own interpretation'

(Image credit: Apple TV Plus)

One of the best Apple TV Plus shows' fanbases have come up with some absolutely fascinating theories since the mystery thriller initially premiered in February 2022. And, while viewers continued to analyze one of 2025's most anticipated shows in the three-year gap between season 1 and its forthcoming sequel, the theory mill has gone into overdrive ever since Severance season 2's first trailer dropped in October 2024. Things escalated further with the launch of Severance season 2's official trailer two months later, too, as fans picked through every second of footage to try and determine what'll happen in the show's sophomore outing.

If you're hoping that season 2 and future installments will confirm or disprove every theory that fans have concocted, though, you'll want to check your expectations. Responding to my query about how Erickson strikes a balance between expanding on or unraveling certain mysteries, and leaving others open to audience interpretation, he said: "It's funny because, as a kid doing my homework, I always made sure to show all of my work so my teacher knew I'd done everything correctly.

We haven't figured everything out, but we're getting there

Dan Erickson, Severance's creator

"Now, because we take great care to intricately figure all of these things out and answer every question [that fans have], my impulse is to put that work on the screen. I want to tell the audience everything. Oftentimes, Ben [Still, director/executive producer] has to be there to put a hand on my shoulder and say 'it's okay to leave some things left unsaid'.

"Part of the fun is letting people speculate and come up with their own interpretation, or they can find their own way to what we were saying and meet us in the middle," he added. "So, I've learned a lot about restraint from Ben, and we still have conversations about which questions we want to answer. We haven't figured everything out, but we're getting there."

Before the highly-rated series returns, read my Severance season 2 review to get an insight into its first four episodes. Once you have, read on for more Severance coverage from me and the wider TechRadar team.

You might also like

• Severance creator 'has a sense' of what the Apple TV Plus show's final scene will be – and how many seasons would be 'ideal' to reach it
• Severance is streaming for free on Roku right now ahead of its season 2 premiere
• Find out what else is coming to Apple TV Plus this January alongside Severance season 2

The basic framework, officials familiar with the talks say, is an exchange of Israeli hostages for Palestinian prisoners, and a six-week ceasefire. Mediators are applying pressure on all parties.

(Image credit: Abdel Kareem Hana)

Three officials said the next days would be critical for ending more than a year of fighting that has shaken the Middle East, but talks have stalled near the finish line repeatedly in the past year.

(Image credit: Abdel Kareem Hana)

Sonos had a very bad 2024, starting with the relaunch of its app in March, which instantly riled up customers with its missing features and sluggish performance. Sonos spent the rest of the year fixing it, costing vast sums of money and reportedly delaying other products as a result. During all this, the company launched its first pair of headphones, which were not a big hit.

A lot of Sonos users turned their ire towards CEO Patrick Spence, and it seems that Sonos board agreed – Spence is leaving Sonos immediately, with a former Pandora (the music-streaming service) exec stepping in to replace him on a temporary basis, while the search for a long-term CEO continues (via Bloomberg).

Tom Conrad is the name of the man taking over for now, and he says he has a prominent arm tattoo of the Sonos Ace headphones, among many other tech tattoos, so we might say that he wears his love of the company on his sleeve.

Conrad said in his letter to employees (via The Verge):

"I think we’ll all agree that this year we’ve let far too many people down. As we’ve seen, getting some important things right (Arc Ultra and Ace are remarkable products!) is just not enough when our customers’ alarms don’t go off, their kids can’t hear their playlist during breakfast, their surrounds don’t fire, or they can’t pause the music in time to answer the buzzing doorbell."

"I’m here to get us back on track. But is getting back on track enough?"

"I think the answer is clearly no. Getting back to basics is necessary, but clearly not enough to unlock the future we all envision for Sonos. So as delighted as I’ll be when every Sonos customer I meet tells me “You work at Sonos!? I love my Sonos!”, what really gets me up in the morning is the idea that we can expand the Sonos platform well beyond “out loud audio at home.”

"I’ve heard from many of you about your own frustrations about how far we’ve drifted from our shared ideals. There’s a tremendous amount of work in front of us, including what I’m sure will be some very challenging moments, decisions, and trade-offs, but I’m energized by the passion I see all around me for doing right by our customers and getting back to the innovation that is at the heart of Sonos’ incredible history."

The Sonos Arc Ultra was the best thing Sonos did in 2024, but it was a low bar to clear, to be fair. (Image credit: Peter Hoffmann)

The letter above appears to reference obliquely several of the issues that led Sonos to such a terrible place, such as restructures that changed how product development was focused, and ignoring growing technical debt, meaning that problems could compound until they became disastrous – you can read far more about Sonos' slow decline internally here.

Which suggests that the new CEO understands the fundamental issues, as well as acknowledging how problems internally lead to problems for the people who buy the products. In particular, the way Conrad raises the idea that people should respond positively to the idea that someone works for Sonos suggests that he's really paid attention to just how severely the company's reputation has been damaged.

It used to have some of the best word-of-mouth in the world of tech, but if you look now at any Sonos-related video on TechRadar's TikTok account, you'll see a sea of comments saying to never buy anything from the company due to the new app (which has, to be fair, improved massively since launch).

So it's good news that the problems appear to have been identified by the interim CEO. The big question is whether he'll have the ability to change them in his time there, and whether the permanent CEO shares his views – or even if all things go right long-term, whether Sonos will ever get its prestige back in the same way.

You might also like

• Sonos Ace review: excellent with movies, awkward with music
• Sonos Arc Ultra review: the best all-in-one soundbar for the price
• Best Sonos speakers: the top buys based on our testing

SteelSeries Apex Pro TKL Wireless Gen 3: one-minute review

The SteelSeries Apex Pro TKL Wireless Gen 3 is the latest incarnation in the brand’s line of elite gaming keyboards, with various improvements for even better gaming performance.

The looks downplay its gaming prowess, with the RGB lighting and floating keycaps being the only signals in this regard. Still, it looks smart and sophisticated, and the build quality is excellent.

It’s quite thick yet not as heavy as you might expect. Every aspect of its built feels premium, right down to the included wrist pad, which is soft yet durable; it’s easily on a par with the best gaming keyboard out there. The double shot PBT keycaps also feel great, with a pleasantly grainy texture.

There are various keyboard shortcuts available by default, accessible via SteelSeries’ Meta modifier key (essentially a function key). These include altering the global actuation point, toggling rapid trigger, changing profiles, recording macros, and switching between three Bluetooth devices.

Similar functions can also be found on the OLED screen in the top right corner of the Apex Pro, with navigation achieved via a dedicated button and the volume wheel. This screen can also be customized in SteelSeries’ GG software, letting you upload images and GIFs to display, and can be used to view system and in-game information.

Overall GG is easy to use and offers plenty of other customizations to sink your teeth into. There are various keyboard shortcuts to configure, although the system-level functions are lacking somewhat. There are also default profiles for certain games, including Counter-Strike 2, DOTA 2, Minecraft and League of Legends. You can set binds to activate on press or release, and the dual actuation mode lets you assign two binds to a single key, each one activated depending on how hard you press, which works effectively. It’s a shame, though, that there’s no visual feedback to show how your actuation settings work in practice.

When it comes to gaming, the Apex Pro really shines: the clack of key presses is very satisfying, the indented keycaps provide a planted finger position, and the bottom row is easy to access thanks to the slope of the keycaps. The layout is quite compact, but I still found it manageable. The switches offer fast yet smooth performance, and hitting certain actuation points is easy thanks to their feel. It is a shame, however, there aren’t more settings available to emulate the analog control of a gamepad.

Also, the wrist rest also cuts off abruptly with no downwards slope, so my palm hung off the edge awkwardly while in the WASD position. Without it, though, I found the keys were raised too high to reach without adopting a more fingertip-oriented position. Typing with the Apex Pro isn’t the best experience either. I personally prefer shallower keycaps, and I found gliding around on the Apex Pro awkward.

There’s no doubt the Apex Pro is a premium piece of hardware, and it commands a premium price tag. It may not be as expensive as some of its rivals, such as the Razer BlackWidow V4 Pro 75%, but it runs close. There are cheaper gaming keyboards with analog switches, such as the Turtle Beach Vulcan 2 TKL Pro, although this is wired only and can’t quite match the Apex Pro for performance. So if you’ve got the money to spend, the Apex Pro is hard to top for an analog keyboard.

(Image credit: Future) SteelSeries Apex Pro TKL Wireless Gen 3 review: price and availability

• $269 / £259 (about AU$456)
• Available now
• Top-end of the market

The Apex Pro costs $269 / £259 (about AU$456) and is available now in black only. It comes with a wrist pad and replacement keycaps can be purchased from SteelSeries’ website. A mini version is also available which loses the OLED screen, and a full-size variant adds a number pad as well as a larger scroll wheel and screen button.

The Apex Pro sits atop the market in terms of price, eclipsing some of Razer’s best offerings, for instance. Although not quite as pricey as the BlackWidow V4 Pro 75%, it’s more expensive than the Huntsman V2 Analog, our pick as the best gaming keyboard. What sets that board apart is the 360-degree motion of the switches keys, letting you mimic an analog stick on a gamepad even more closely. It’s wired only, though.

If you’re looking for another analog keyboard, then the Turtle Beach Vulcan 2 TKL Pro is considerably less expensive than the Apex Pro. Although it’s wired only and has some issues with its analog implementation, the Vulcan 2 is still a worthy contender in the sector.

Value: 3 / 5

(Image credit: Future) SteelSeries Apex Pro TKL Wireless Gen 3 review: specs

(Image credit: Future) SteelSeries Apex Pro TKL Wireless Gen 3 review: design and features

• Premium design
• Useful OLED screen
• Intuitive software

The Apex Pro looks understated for a gaming keyboard. The totally black design is bereft of the brash trappings common in the sector, so at first glance you’ll struggle to identify its intended audience. The floating keycaps, though, are a giveaway, as is the opulent RGB lighting.

There’s no denying the Apex Pro has a premium aura, though. It’s also substantial, with a thick, two-tiered chassis, but fortunately it isn’t as heavy as it looks. The build quality is impeccable, with every material, join and key feeling smooth and sturdy.

The double-shot PBT keycaps are lightly textured and feel great under your fingers. They’re quite thick and sit high above the board, allowing for deep presses. If they aren’t to your liking, though, they can be swapped out, and a puller is included. The layout is also quite compact, which I had to get used to, but it didn’t take long to adapt.

The wrist pad is similarly well-built yet pleasingly light, with a smooth rubberized layer that’s both durable and comfortable. It connects magnetically to the board itself, and although the connection is quite weak, it suffices.

There are several default hotkeys on the Apex Pro, accessible with SteelSeries’ own modifier key, called the Meta key. These include global adjustments for the actuation point and rapid trigger function, as well changing profiles, recording macros, and switching between three Bluetooth devices.

However, more detailed settings and customizations are available via the Engine tool within the GG software. All keys can be remapped, and can be set to trigger on press or release. There’s a good selection of actions, including profile switching, keyboard setting adjustments and app launching. There are also OS shortcuts that can be mapped, including a selection of media keys and copy and paste, but there aren’t many more than this; those in other peripheral software are more numerous.

(Image credit: Future)

GG also allows you to set actuation points ranging from 0.1mm to 4mm for all main keys. You can also toggle rapid trigger for each individual key, which, as GG explains, dynamically alters the actuation point when spamming, allowing for rapid-twitch movements. It even has a dual binding mode, allowing a single key to have two different binds, depending on the primary and secondary actuation points of your choosing.

For the most part, GG is easy to use with a clear layout and some explanations for various settings. However, it would’ve been nice to see some visual feedback when adjusting actuation points and rapid trigger modes, as some of the Apex Pro’s rivals offer with their own software.

Nestled in the top right corner of the Apex Pro is a small monotone OLED screen, called the Smart Screen. It displays battery life by default, but can be customized in GG to display anything from your own images and GIFs, to system and in-game information.

The options available for the latter are quite in depth. For instance, you can choose to display certain information at certain increments, and choose various screens to cycle through, each displaying their own unique information at various points. Supported games include Counter-Strike 2, DOTA 2, Minecraft, and League of Legends.

There’s also a button on the screen that can be used to control media playback, such as play / pause and next / previous functions. You can enter a menu by holding this button down for a few seconds, from which more settings and functions are available: you can change the brightness of the RGB illumination, toggle between three of your connected Bluetooth devices, and change the global actuation point and rapid trigger functions. Navigation in this menu is achieved with the volume wheel acting as a scroller, and it’s responsive and effective.

Design & features: 4 / 5

SteelSeries Apex Pro TKL Wireless Gen 3 review: performance

• Superb gaming performance
• Analog keys actually usable
• Typing not the best

The gaming performance of the Apex Pro is excellent. The stock keycaps are nicely indented, allowing you to maintain a secure position, and presses are weighted just right, allowing for easy actuation that provides pleasant feedback. Reaching the CTRL key and spacebar is also easy, thanks to the slight incline of the keycaps.

The switches are silky smooth and well lubed, offering plenty of clack without being too loud. I also found it relatively easy to hit certain actuation points, as the Hall-effect magnetic switches have more feel and resistance than others I’ve tried, such as those on the Turtle Beach Vulcan 2 TKL Pro. This makes the Apex Pro ideal for rapid trigger functionality or dual binding.

SteelSeries has also implemented a new Protection mode, which prevents keys adjacent to the intended one from accidentally triggering by reducing their sensitivity momentarily, which is a useful feature.

Where it does miss a trick, however, is the ability to map the keys to an analog stick. This would make for smoother character movements and vehicle control, and is something that can be done with the aforementioned Turtle Beach Vulcan 2 TKL Pro, as well as the Razer Huntsman V2 Analog.

(Image credit: Future)

As someone who prefers shallow keycaps for typing, I didn’t find the Apex Pro well suited for productivity purposes. Although adjusting the actuation point can help to prevent typos, the relatively thick keycaps and their deep indentations aren’t conducive to gliding around the board.

The included wrist pad is pretty much essential for typing comfortably, as the thick body forces your wrists to bend awkwardly without one. Thankfully, the one included offers enough support and padding to make both typing and gaming more comfortable.

The wrist pad can be more cumbersome when gaming, though. As the backend tails off abruptly with no slope, the area where your wrist meets your palm tends to overhang when in the WASD position, which is a strange and not entirely comfortable sensation. I did find that I could game without the pad, but I had to adapt my fingers so they were more perpendicular to the key caps, emphasizing more contact with the fingertips. If this is your preferred playing style anyway, then you shouldn’t have a problem.

Steelseries claims a battery life of up 40 hours over Bluetooth, or 37.5 hours via the 2.4GHz dongle. During my tests I found this to be accurate, as I was able to get several days use out of it.

Performance: 4.5 / 5

Should I buy the SteelSeries Apex Pro TKL Wireless Gen 3? Buy it if…

You want analog keys
Analog keys aren’t all that common, so if you want changeable sensitivity and rapid trigger functionality, this could be the board for you.

You want top-tier performance
Gaming with the Apex Pro is sublime, thanks to its smooth and fast performance. Actuation points are easy to hit and the rapid trigger works well.

Don’t buy it if…

You’re on a budget
You’ll be hard pressed to find many gaming keyboards that perform better than this, but it’ll cost you.

You want controller emulation
Although the switches are analog, you can’t emulate a gamepad’s triggers or analog sticks as you can with some other analog boards.

SteelSeries Apex Pro TKL Wireless Gen 3 review: also consider

Razer Huntsman V2 Analog
If you must have an analog keyboard, then the Huntsman V2 Analog is another fine example in this category. Razer claims that the analog inputs can work with any controller-based game, allowing for ultra-smooth inputs. Uniquely, the keys also feature 360-degree movement, emulating a joystick more closely. The Huntsman V2 Analog can be had for slightly less than the Apex Pro, although if you hunt around you may be able to find a better deal. Read our Razer Huntsman V2 Analog review.

Turtle Beach Vulcan 2 TKL Pro
If you still want analog switches but want to save even more cash, then the Vulcan 2 TKL Pro could fit the bill. It also lets you emulate the triggers and analog sticks of a gamepad, although the implementation isn’t as seamless as you might imagine. Still, it can have its uses. Read our Turtle Beach Vulcan 2 TKL Pro review.

How I tested the SteelSeries Apex Pro TKL Wireless Gen 3

• Tested for about a week
• Used for gaming, productivity and general use
• 10+ years PC gaming experience

I tested the Apex Pro for about a week. During that time I used it for gaming, productivity and general use.

I played various games, including Counter-Strike 2 and I Am Your Beast, which were useful for testing the actuation points and rapid trigger functionality. I also made sure to make use of every feature and function available via the GG software where possible.

I have been PC gaming for over 10 years, and have experienced numerous keyboards during that time, spanning various switch types, key profiles, and form factors. I have also reviewed multiple gaming keyboards from a mixture of brands.

• First reviewed November 2024
• Read more about how we test

• Two Hatsune Miku Fortnite skins have leaked ahead of their official reveals
• They will reportedly be available tomorrow
• One is apart of the game's Festival Pass, while the other is in the Item Shop

Virtual pop star and Vocaloid synthesizer software icon Hatsune Miku is reportedly coming to Fortnite tomorrow in the form of two new skins. Both have leaked early via social media - but they look a little bit off to me.

Images purportedly showing the upcoming Hatsune Miku x Fortnite collaboration have been doing the rounds on social media this morning. They suggest that two skins of the character will be added to the game tomorrow: Classic Miku and Neko Miku.

Classic Miku will apparently be available via the Item Shop and seems to depict Miku in her classic baby blue ensemble. In addition to being able to buy the skin on its own, the leaks indicate that it will be bundled with a cute Miku-themed rucksack Back Bling and a complimentary trail.

The Neko Miku skin will then be part of the Fortnite Festival Music Pass, a premium battle pass centered around the new spinoff rhythm game mode. If you're keen to get your hands on this one, leakers are indicating that the season will end on April 8.

On top of all this, it seems that at least three new instruments and a Jam Track for the song Magical Cure Love Shot! will also be added.

What did they do to her?

Although I'm always up for some Hatsune Miku rep, I can't pretend that I'm not a little disappointed by the look of these skins so far. As you might expect, they use the cel-shaded Fortnite anime art style that we've already seen used for many other skins, but something about them just seems a bit wrong.

She has a weird, almost blank look and her eyes seem not only too small but also look just a little too far apart to me. On the whole, while they definitely don't seem like the worst skins ever, I think that they could easily be made a little cuter through some subtle tweaks.

I reckon a design more closely based on her appearance in the game Hatsune Miku: Project DIVA Mega Mix, which boasts a similar cel-shaded style, would have been a much better choice and a more accurate representation of how she looks in general.

I'm also really hoping that Magical Cure Love Shot! isn't the only Jam Track on the cards. There are countless more iconic Miku songs out there, like World is Mine or Romeo and Cinderella, that I would much rather play.

That said, I can't pretend that I won't still be buying both this skin bundle and Festival Pass instantaneously if they do indeed drop as per current leaks. Fingers crossed everything looks better in-game than in these leaked pics.

You might also like...

• Nintendo has finally commented on the Switch 2 leaks, but it's only left me with more questions
• Could Halo and Microsoft Flight Simulator come to Nintendo Switch 2 and PS5? Yes, per latest rumors
• These new Hyperkin Nintendo Switch accessories have me seriously hyped as an anime lover

• CIOs and CTOs predict losing 3% of their workforce in the next 3-5 years
• Some say they may see job losses equating to 5-10%
• Banking profits could skyrocket from boosted productivity

A new Bloomberg report has claimed as many as 200,000 jobs on Wall Street could be lost over the next three to five years as artificial intelligence becomes more integrated into workflows.

The data comes from a Bloomberg Intelligence survey, which found Chief Information and Technology Officers are expecting to cut 3% of jobs over the next few years. Nearly one in four predicted job cuts of between 5-10% of their headcounts.

Senior analyst and report writer Tomasz Noetzel added back office, middle office and operations roles are most at risk of displacement – essentially, the roles involve routine and repetitive tasks that can be automated for improved efficiency.

AI is set to cost thousands of banking jobs

Although AI and automation is threatening to take work away from humans, Noetzel said, “AI will not eliminate them fully, rather it will lead to workforce transformation.”

Boosted efficiency could see significant cost reductions for banks – the report suggests pre-tax profits could be 12-17% higher in 2027 than they are today. Around four in five see AI increasing productivity and revenue generation by at least 5% in the next three to five years.

However, Bloomberg Intelligence’s report seems conservative when compared with other reports. Last summer, Citi said that more than half (54%) of banking jobs are at high risk of being automated.

While the outlook may seem negative, a separate report from the World Economic Forum reveals that all hope is not lost. By the end of the decade across all sectors, WEF says we could see a net increase of 78 million jobs, even though 92 million workers could be displaced.

That same WEF report echoed Bloomberg Intelligence’s findings, revealing that bank tellers are among the most at-risk workers.

You might also like

• AI jobs are among the fastest-growing in the UK
• Check out the best job sites and best recruitment platforms
• Enhance your upskilling efforts by logging on to the best online learning platforms

• The UK government has revealed a 50 point plan to inject AI into the public sector
• Plan comes with a £14 billion private sector investment
• Prime Minister addresses safety concerns, but doesn't want the UK to 'fall behind'

The UK Government has laid out plans to push AI into all corners of British industry with a multi-billion pound investment.

The Labour Government plans to ‘unleash’ AI and make the UK a ‘world leader’ by accepting £14 billion in private sector investment, and has revealed a 50 point plan that will see AI ‘mainlines into the veins’ of public services.

This includes a plan to ‘unlock’ public data by handing it over to ‘researchers and innovators’, including anonymized NHS data, which will be made available to train AI models. The government says there will be “strong privacy-preserving safeguards” and the data would never be owned by private companies.

Creating jobs

The plans will also look to free up time for public sector workers by allowing AI to undertake admin tasks, and will introduce ‘AI Growth Zones’ which will speed up AI infrastructure, rapidly building data centers and giving them better access to the energy grid - despite environmental concerns.

Private sector investment comes primarily from Vantage Data Centres, a firm which plans to invest £12 billion in data centre development, which will create over 11,500 jobs in the process. Also investing are Nscale and Kyndryl, two leading digital companies who plan to help roll out AI into Britain’s economy.

‘’The UK Government’s AI Opportunities Action Plan will play an important role in helping the UK to unlock the full potential of AI,” said Alison Kay, Vice President, UK and Ireland at Amazon Web Services

“By putting close industry collaboration and public-private partnership at the heart of the Government’s agenda, every citizen, community and business in every region of the UK will have the opportunity to realise the benefits of AI, and thrive and grow’’

A disruptive technology

The Prime Minister says that AI can ‘transform the lives of working people’, speeding up planning consultations, driving down admin, and feeding AI through road cameras to help spot potholes - although oddly, my car can spot potholes just fine.

Minister Pat McFadden said AI is a ‘test and learn’ technology, and, "mistakes will be made" with the applications, but wouldn’t confirm whether this plan will mean short-term job losses for workers. Addressing objections to the plan, the Minister warned if the UK doesn’t develop the technology here, then the country would ‘lose out’ to other nations in the AI race.

In the Prime Minister's speech announcing the news, he acknowledged the anxieties surrounding AI, but affirmed the real danger is in falling behind other nations, noting, "the far bigger risk is we don't go for it," and miss out on 'massive' opportunities.

AI makes services more human

The Prime Minister also outlined the ways in which AI can help make our public services 'more human', primarily referring to healthcare and social work, in the speech. By taking administrative and mundane tasks away from public sector workers, services will be 'more human' by allowing staff to reconnect with patients and those under their care.

Starmer promised to "make sure this technology is safe" by building on the previous government's 'world-leading' AI safety institute, promising to uphold and develop safeguarding measures to ensure AI is used safely.

"Change on this scale and at this speed can be worrying, especially when experts warn of the safety risks,"the Prime Minister said.

A push for productivity

Productivity in Britain has been somewhat stagnant for the last 15 years, but by using the full power of AI, the Prime Minister hopes to give the British economy the boost it so desperately needs.

That is, of course, if AI can increase efficiency for workers, rather than result in widespread unemployment.

Recent research suggests AI could displace up to 275,000 jobs per year at the peak of adoption. Politicians and tech leaders alike have long insisted AI will simply take over the mundane tasks from workers, rather than replace them, but as someone who has worked one of the many jobs entirely made up of 'mundane' tasks, I have to be skeptical.

You might also like

• Take a look at our pick of the AI tools around
• AI jobs are among the fastest-growing in the UK
• Check out our choices for best antivirus software

If you are looking for a printer that is quick, efficient and easy to use, these are some of the best options in the market.

Santa Ana winds are expected to pick back up, threatening the containment of Southern California wildfires. And, Homeland Security Secretary Alejandro Mayorkas reflects on his career.

(Image credit: Eric Thayer)

Which type of RAM should you buy? Here are the differences explained in nerdy detail.

• Some Docker Desktop users on Mac are unable to open their application
• A false malware alert is being shown on startup
• Users can fix the issue with a patch, so download now

Docker has confirmed some versions of its Docker Desktop software are failing to start or work properly as a result of a false malware warning.

Posting to GitHub, Senior Principal Engineer Christian Dupuis shared a screenshot of the pop-up, which reads: “com.docker.vmnetd was not opened because it contains malware.”

Dupuis confirmed users seeing this message are unlikely to be affected by malware, with the cause of the problem already identified and a fix in the works.

Docker Desktop malware warning

“The reason for this warning is that some files in existing installations are incorrectly signed," the post added.

Described as a “one-click-install application for your Mac, Linux, or Windows environment that lets you build, share, and run containerized applications and microservices,” Docker Desktop will need updating to fix the problem.

Version 4.37.2 is the latest release at the time of writing, which promises to address the issue. Docker says users should upgrade through the app where possible, but a set of instructions are available for those unable to bypass the malware warning message.

A separate patch is available for versions 4.32 - 4.36. Versions 4.28 and earlier are unaffected by the error.

Speaking about version 4.37.2, Dupuis said: “This version prevents installations from entering into the failure state. Please update to this version as soon as possible to prevent future issues.”

A status page updated recently confirms Docker Desktop is still experiencing partial service disruption.

“We know how critical Docker Desktop is to your workflow and are committed to ensuring a smooth resolution," Director of Engineering Chris Crone added.

"A temporary workaround that will restore functionality is available for any affected users… Our team is prioritizing this issue and working diligently on a permanent fix. If you prefer to wait for the longer-term patch update, please refrain from (re)-starting Docker Desktop.”

You might also like

• This devious macOS malware is evading capture by using Apple's own encryption
• Downloaded something dodgy? Here’s the best firewall options to stay safe
• Consider using the best endpoint protection software