TechRadar News

Following the recent resignation of one of Anthropic’s leading researchers, multiple AI CEOs have suddenly begun calling for a slowdown in the development of AI technology to allow regulations and governance on the technology to catch up.

Speaking to the BBC after his resignation, Jacob Coxon warned, “I believe that if we don't slow down at the current rate of progress, there is a strong chance that we could all die in the immediate future.”

Following this, Anthropic head Dario Amodei, OpenAI CEO Sam Altman, and Grok founder Elon Musk have all apparently aligned in their calls for development to slow down. But there are some tricky waters to navigate - particularly around President Trump, China, and what guardrails should be put into place.

What are AI heads saying?

Over the weekend, Amodei posted an essay on “why the AI industry should slow down”. In it, he said, “I believe that if slowing down bought us even an extra year or two before models reach critical levels of capability, and we used that time to advance alignment, we could greatly reduce the risk that something goes seriously wrong.”

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.You can read the full post here: https://t.co/OGyPb7yaYtSeptember 12, 2026

Within the essay, Amodei outlined how AI could be ‘paced’ within the US, and globally, alongside a recommendation that AI companies put ‘evaluator’ teams into place to ensure AI models stay aligned to their tasks. Elon Musk replied to Amodei’s social media post, stating that the Anthropic head was “right”.

Sam Altman told Fortune the regulations and standards for AI further were “not at a place” to continue progressing AI development.

(Image credit: Future)

Got an opinion for us? Here’s how you can submit your perspective

But not everyone is convinced. US President Donald Trump has said that slowing down AI development is non-negotiable, as it would allow China to rapidly catch up to US AI capabilities. He told reporters that the US is “leading China on AI... and, frankly, I want to keep it that way,” adding that “whoever wins AI, wins”.

Trump also said that “very negative forces” were behind the growing opposition to AI, and said that fears were being stoked by “that won’t happen”.

China also isn’t convinced by what the AI giants are saying. Beijing labelled the calls for a slowdown as “fearmongering” from a “Cold War playbook.” In his essay, Amodei said that a “Chinese lead in AI would pose grave danger for the United States and the world.”

Chinese Foreign Ministry spokesperson Guo Jiakun said, “Fearmongering, confrontation and malicious competition will only disrupt the process of global AI governance and serve no one’s interests.”

Expert perspectives on calls for AI slowdown

• John Strand, Owner, Black Hills Information Security:

Up until this weekend, I was leaning toward believing that calls for an AI slowdown were purely performative. Then I woke up and read the news today and realized that it almost doesn’t matter.

We can talk about slowing down AI until we’re blue in the face, but when the United States is saying it doesn’t want to slow down because it’s competing with China, and China is aggressively pushing AI development as well, we’re talking about the two major economic and military powers on the planet having enormous incentives to keep moving.

This really feels like we’re entering an atomic arms race moment.

At that point, calls for a slowdown don’t have much bite.

I’d like to believe that Anthropic, OpenAI, xAI, and the other frontier model labs are working on better controls. But unless you can get the nation states and the major AI labs moving in the same direction, I don’t see how meaningful restrictions actually work.

This really feels like we’re entering an atomic arms race moment.

Stick with me here.

In 1950, physicist Leó Szilárd publicly discussed the idea of a cobalt bomb, essentially a doomsday weapon that could potentially produce enough radioactive fallout to make the Earth uninhabitable. He wasn’t proposing that somebody build the damn thing. He was trying to demonstrate where the technology could ultimately lead.

That’s the kind of moment I think we’re approaching with AI.

During the nuclear arms race, eventually the consequences became serious enough that competing nations had to at least start talking about limits, controls, and ways to keep competition from ending catastrophically.

I think we’re heading toward a similar problem with AI. Until China, the United States, and the major frontier model labs are all sitting at the same table, restrictions adopted by individual companies or individual countries are going to have a very difficult time holding.

Someone slowing down only works if they believe the other guy is going to slow down too.

• Ryan McCurdy, VP, Liquibase:

Slowing frontier development may give AI companies more time to understand and address the risks Amodei is describing. But enterprises can’t build their AI strategy around the assumption that AI is going to slow down.

AI is already moving from generating content and code to taking action across software delivery and production systems. The question for enterprises is how they adopt that capability without giving up control.

We can debate how quickly the frontier should move. Enterprises still have to prepare for where it’s going.

That means putting governance where AI decisions become real actions. Organizations need to define what an agent can access, what it can change, what it can decide on its own, and what policies have to be met before a change reaches a critical system. Those controls need to work whether the action comes from a developer, automation, or an AI agent.

We can debate how quickly the frontier should move. Enterprises still have to prepare for where it’s going.

• Tristan Watkins, director of services innovation, Advania UK:

Until recently, the major AI labs have been reluctant to slow their development efforts unilaterally. Over the last week this changed, with new commitments from OpenAI and Anthropic to prioritise AI alignment and interpretability research, to become more externally verifiable, and to establish safety precedents that governments could adapt.

Hopefully this underscores why we need governments to lead these efforts more proactively.

Given that these two organisations already allocate far more on AI Safety than their competitors, this bilateral leadership is extremely welcome.

It appears that other US labs may follow suit, but given the differences in AI Safety spending outside of Anthropic and OpenAI today, this will require investment more than lip service. Hopefully this underscores why we need governments to lead these efforts more proactively.

• Oleksandr Yaremchuk, CTO and Co-Founder, Manifold Security:

Pacing the frontier is the right conversation to be having, but it cannot become a substitute for securing the AI we have already put into the world. The uncomfortable reality is that we are debating how to quickly build more powerful agents while struggling to control the ones already operating with real credentials, real access and real-world consequences.

The incidents behind this debate make that clear. The Hugging Face attack was not just a failure of model alignment. Agents ran for days through an unmonitored system, with credentials that had not been rotated, and the victim spotted the activity before the people running the agents did. The problem wasn't simply what the model was capable of. It was that nobody was watching closely enough when it acted.

But if an agent can act autonomously on your systems today, you should already be able to answer three basic questions: what did it do, what did it have access to, and could you have stopped it?

A fitting analogy is with hazardous materials. We don't just wait for them to become more dangerous before deciding how they should be handled. We control their custody, monitor where they go, limit who can access them and establish clear accountability when something goes wrong. AI agents need the same thinking.

Independent evaluation of frontier models is important. But if an agent can act autonomously on your systems today, you should already be able to answer three basic questions: what did it do, what did it have access to, and could you have stopped it? If you don't know what it did or what it could access, you can't know whether you could have stopped it. Slowing down the next generation won't solve the problem you have right now.

• Heath Mullins, Chief Evangelist, ExtraHop:

AI leaders calling for a slowdown is confirming what the security industry has already been living through firsthand. This isn't a hypothetical risk, it's the threat landscape we're defending against right now.

While it is concerning to see the pace of innovation behind these AI models, the real challenge is that organizations haven't had the runway to build the infrastructure to defend against machine-speed threats.

This isn't a hypothetical risk, it's the threat landscape we're defending against right now.

Calls for caution surrounding the speed of AI development buys the security industry time to get proper visibility into AI activity.

Understanding AI activity within an organization is critical as we’ve seen models break out of sandboxes despite governance built into those models. Every organization will be relying on AI agents for machine-speed defense, and they need their own governance over how these models and agents operate inside their environment, starting with independent evidence of what they actually do, what they access, where they move data, what systems they talk to, and what actions they take.

You can't govern AI based on what a model is designed or permitted to do. Instead, you need real-time evidence of what models and agents are actually doing, because the gap between exponentially more capable AI and defenders' ability to see it is exactly where the next incident happens.

• Bri Frost, Director of Product Management, Cloud Range:

The answer is not necessarily to stop AI innovation but, we need to stop pretending innovation and security are advancing at the same speed.

When ChatGPT became publicly available in 2022, the models were dramatically less capable than they are today — and the guardrails were very easy to manipulate. The difference is that the models behind those guardrails are no longer the models of 2022. They can reason better, write and debug code. They can operate as agents. They can collaborate! And increasingly, they can interact and affect real infrastructure.

The faster we build the engine, the more important the brakes become.

Meanwhile, the model release cycle has gone from feeling like major capability jumps every year or two to seemingly every few weeks. That creates a dangerous asymmetry: AI capability is compounding faster than security.

Security and innovation have always been in conflict with each other. If every security problem had to be solved before we innovated, we’d never ship anything. But the opposite extreme is just as reckless: accelerating capability while just assuming we’ll bolt the security controls on afterward and they’ll be effective.

Every new release of AI capability expands the attack surface exponentially. Give a vulnerable model better reasoning, then tool access, then memory, then autonomy, then connectivity to production systems, and yesterday’s jailbreak isn’t just a clever prompt anymore — it’s an execution path. That’s the snowball effect we should be worried about.

Responsibility also must lie with the AI companies. If a SaaS company knowingly shipped software with weak security controls and customers were harmed, we wouldn’t excuse it because they were 'innovating quickly'.

So why are we treating AI differently?

You don’t get to race to build increasingly powerful, autonomous systems, profit from them, and then shrug when predictable security failures cause damage.

Sure the argument can be made that no product is perfectly secure - That’s not the standard. But if you ship the product, you inherit responsibility for securing it. And continuing to secure it better!

The conversation shouldn’t simply be “Should we slow AI down?”

It should be: Can our ability to test, validate, contain and secure AI keep pace with our ability to make it more powerful? Is there an equivocal kill switch?

Right now, the answer is no.

And if we’re going to keep accelerating — which I believe we will — then independent testing, adversarial evaluation, isolated testing environments, containment, continuous validation and security-by-design can’t remain optional steps we add after the innovation happens.

The faster we build the engine, the more important the brakes become.

• Denis Calderone, CTO, Suzu Labs:

Amodei's diagnosis is the most honest thing a frontier lab CEO has said publicly. The agent risk is real, recursive self-improvement is accelerating, and the competitive pressure is making both worse.

Where I get skeptical is the prescription. Democratic coordination among companies in a commercial race? Global pacing agreements with China? Amodei himself rates the hardest steps as unlikely. No lab has named a single model release they'll delay because of this essay.

No lab has named a single model release they'll delay because of this essay.

The one idea worth holding the industry to is embedded evaluators with independent publication rights. Give third-party safety researchers permanent access inside the labs, comparable to what bank examiners have inside banks, and let them publish what they find without the company controlling the narrative. That's a simple, concrete accountability mechanism. It doesn't require global coordination or antitrust waivers. Anthropic says they're committing to it unilaterally. Good. Now make the rest of the industry match.

• Donald McFarlane, Board Member, Xcape Inc:

AI does not develop an agenda; its operators do. When we give an autonomous system powerful access and ability to act at machine speed, they will continue to prove highly capable.

AI does not develop an agenda; its operators do.

Rules enacted in the name of safety must not become a moat against competition or progress. Enormous compliance costs may be manageable for the handful of companies already spending billions building frontier models, while becoming a substantial barrier to everyone behind them.

Government can help clarify accountability and duties of care, and facilitate strong information sharing and collective defense, which is an area where we sorely need more effective public-private partnerships.

But safeguards should focus on how these systems are used and deployed, rather than deciding who is allowed to build powerful AI in the first place.

The goal should be safer deployment without pulling up the drawbridge on innovation.

How do I submit my own perspective on emerging news?

If you have an expert perspective you would like to share on an emerging story or particular topic, please get in contact here: benedict.collins@futurenet.com

• Amazon is still selling fake 'energy-saving' plugs despite having pledged to remove them
• Which? first spotted them in 2025, and Amazon said they had been removed
• The retailer promised to remove them for a second time in August, but they're still available to purchase

Fake ‘energy-saving’ plugs are one of the biggest tech scams, and they’re still being sold on Amazon even though the retailer promised MPs it would remove the product listings.

Liberal Democrat MP Joshua Reynolds highlighted the dangers of the plugs — which can pose a fire risk — during a hearing back in June, calling for the retailer to take accountability.

In August, Amazon’s UK Director of Public Policy, Monica Ariño, issued a letter to the House of Commons Business and Trade Select Committee saying the devices had been removed from sale, adding that there was "a technical issue in [our] automated classification system that was preventing some of these products from being proactively detected".

This isn’t the first time Amazon has promised to remove fake plugs. Back in 2025, Which? first spotted these ‘energy-saving’ plugs and approached retailers with their findings, resulting in their removal from stores. However, identical products were found on Amazon just weeks later.

(Image credit: Future)

Despite Amazon's most recent pledge to remove the plugs from sale, they're still available to purchase from the retailer. We’ve checked Amazon UK’s page and found that these products are still live — the Telegraph (£) has also spotted similar listings.

Now, Reynolds says he’ll be writing to Amazon asking for an explanation, adding that "Amazon has got some proper questions they have to answer. These are exactly the same products tested by Which?". He also claims that these plugs are appearing as sponsored ads at the top of Amazon’s listing pages.

As mentioned, these devices were first spotted during a test carried out by Which?, for where it tested eight devices from sites including eBay, AliExpress, Shein, and TikTok Shop. It found no evidence that the plugs actually worked, and noted that it would be illegal to sell them in the UK, as they didn't meet basic safety standards.

Which? found that each plug it tested had ill-fitted socket pins (Image credit: Which? )

So how are these plugs claimed to cut your energy bills? Based on the product listings we’ve found, Amazon’s ‘energy-saving’ plugs are advertised as ‘voltage stabilizers’ and ‘surge protectors’, and are claimed to significantly cut your energy bills by optimizing the performance of your household appliances.

At a time when energy bills are increasing, it’s an easy sales pitch to make, and they’re also affordably priced (usually around the £10-£15 mark), making them seem to be a worthwhile puchase.

However, it's claimed that the devices are a fire hazard, and also pose an electrocution risk. According to the tests conducted by Which?, more than half of the plugs’ socket pins were either too small, not straight, or were so frail they just snapped off.

Other tests found that some of the plugs failed to meet basic safety measures due to poor manufacturing, including non-standard capacitors and soldering.

Amazon hasn't yet commented on the current status of these product listings, or responded to the latest claims. We've asked the company for a comment, and will update this story if we get a response.

• Amazon is still selling fake 'energy-saving' plugs despite having pledged to remove them
• Which? first spotted them in 2025, and Amazon said they had been removed
• The retailer promised to remove them for a second time in August, but they're still available to purchase

Fake ‘energy-saving’ plugs are one of the biggest tech scams, and they’re still being sold on Amazon even though the retailer promised MPs it would remove the product listings.

Liberal Democrat MP Joshua Reynolds highlighted the dangers of the plugs — which can pose a fire risk — during a hearing back in June, calling for the retailer to take accountability.

In August, Amazon’s UK Director of Public Policy, Monica Ariño, issued a letter to the House of Commons Business and Trade Select Committee saying the devices had been removed from sale, adding that there was "a technical issue in [our] automated classification system that was preventing some of these products from being proactively detected".

This isn’t the first time Amazon has promised to remove fake plugs. Back in 2025, Which? first spotted these ‘energy-saving’ plugs and approached retailers with their findings, resulting in their removal from stores. However, identical products were found on Amazon just weeks later.

(Image credit: Future)

Despite Amazon's most recent pledge to remove the plugs from sale, they're still available to purchase from the retailer. We’ve checked Amazon UK’s page and found that these products are still live — the Telegraph (£) has also spotted similar listings.

Now, Reynolds says he’ll be writing to Amazon asking for an explanation, adding that "Amazon has got some proper questions they have to answer. These are exactly the same products tested by Which?". He also claims that these plugs are appearing as sponsored ads at the top of Amazon’s listing pages.

As mentioned, these devices were first spotted during a test carried out by Which?, for where it tested eight devices from sites including eBay, AliExpress, Shein, and TikTok Shop. It found no evidence that the plugs actually worked, and noted that it would be illegal to sell them in the UK, as they didn't meet basic safety standards.

Which? found that each plug it tested had ill-fitted socket pins (Image credit: Which? )

So how are these plugs claimed to cut your energy bills? Based on the product listings we’ve found, Amazon’s ‘energy-saving’ plugs are advertised as ‘voltage stabilizers’ and ‘surge protectors’, and are claimed to significantly cut your energy bills by optimizing the performance of your household appliances.

At a time when energy bills are increasing, it’s an easy sales pitch to make, and they’re also affordably priced (usually around the £10-£15 mark), making them seem to be a worthwhile puchase.

However, it's claimed that the devices are a fire hazard, and also pose an electrocution risk. According to the tests conducted by Which?, more than half of the plugs’ socket pins were either too small, not straight, or were so frail they just snapped off.

Other tests found that some of the plugs failed to meet basic safety measures due to poor manufacturing, including non-standard capacitors and soldering.

Amazon hasn't yet commented on the current status of these product listings, or responded to the latest claims. We've asked the company for a comment, and will update this story if we get a response.

Welcome to our live coverage of Dreamforce 2026!

We're on the ground in San Francisco and ready to cover all the big news and events happening this week.

We'll see keynotes from CEO Marc Benioff, likely joined by a number of special guests, along with new product announcements and much more - so stay tuned to TechRadar Pro for all the latest updates.

When is Dreamforce 2026?

Dreamforce 2026 is taking place Tuesday 15th - Thursday 17th September 2026.

What time does Dreamforce 2026 start?

Dreamforce 2026 kicks off at 10am PST on Tuesday 15 September, when CEO Marc Benioff starts things off with his usual opening keynote.

Where is Dreamforce 2026?

Dreamforce 2026 is taking place at the Moscone Conference Center (all of it, that's how big it is) and a variety of other downtown locations in San Francisco, California.

Can I attend Dreamforce 2026?

You can - but be quick! The last few tickets are selling fast.

Can I watch Dreamforce 2026 online?

You can indeed - the company is streaming much of the event on its Salesforce+ platform, where they will be a dedicated Dreamforce+ section showing 400+ sessions, virtual hands-on trainings, and more.

If you aren't already signed up, you can register for Salesforce+ for free today.

What should we expect?

One of the best things about Dreamforce for us is that there's always so much going on - lots of news, product announcements and more.

From the Salesforce side, we're all set to hear about its latest and greatest AI announcements and updates. Last year was all about AI agents, and the company is now set to reveal more changes to Agentforce 360 which can help make workers more productive than ever.

As for Slack, it recently announced Slackforce, promising deeper connection with the Salesforce platforms, with AI agents now able to do more than ever before - so we'll doubtless hear more about that.

And alongside that, you've got Tableau, Mulesoft, and all of the company's other arms, all of which will doubtless have new things to shout about.

There's always room for something a little extra with Salesforce too - whether that's the traditional Ohana greeting, Matthew McConaghey-voiced intro videos, celebrity cameos and musical guests, it always keeps us on our toes.

September 14, 2026 – 10:44 AM

Hello from San Francisco! We're excited for a week of Dreamforce, so stay tuned for all the latest updates as they happen.

The event kicks off tomorrow (Tuesday 15) so we'll try and get a sneak peek at the site (affectionally known as the Campgrounds) and anything else we spy.

Welcome to our live coverage of Dreamforce 2026!

We're on the ground in San Francisco and ready to cover all the big news and events happening this week.

We'll see keynotes from CEO Marc Benioff, likely joined by a number of special guests, along with new product announcements and much more - so stay tuned to TechRadar Pro for all the latest updates.

When is Dreamforce 2026?

Dreamforce 2026 is taking place Tuesday 15th - Thursday 17th September 2026.

What time does Dreamforce 2026 start?

Dreamforce 2026 kicks off at 10am PST on Tuesday 15 September, when CEO Marc Benioff starts things off with his usual opening keynote.

Where is Dreamforce 2026?

Dreamforce 2026 is taking place at the Moscone Conference Center (all of it, that's how big it is) and a variety of other downtown locations in San Francisco, California.

Can I attend Dreamforce 2026?

You can - but be quick! The last few tickets are selling fast.

Can I watch Dreamforce 2026 online?

You can indeed - the company is streaming much of the event on its Salesforce+ platform, where they will be a dedicated Dreamforce+ section showing 400+ sessions, virtual hands-on trainings, and more.

If you aren't already signed up, you can register for Salesforce+ for free today.

What should we expect?

One of the best things about Dreamforce for us is that there's always so much going on - lots of news, product announcements and more.

From the Salesforce side, we're all set to hear about its latest and greatest AI announcements and updates. Last year was all about AI agents, and the company is now set to reveal more changes to Agentforce 360 which can help make workers more productive than ever.

As for Slack, it recently announced Slackforce, promising deeper connection with the Salesforce platforms, with AI agents now able to do more than ever before - so we'll doubtless hear more about that.

And alongside that, you've got Tableau, Mulesoft, and all of the company's other arms, all of which will doubtless have new things to shout about.

There's always room for something a little extra with Salesforce too - whether that's the traditional Ohana greeting, Matthew McConaghey-voiced intro videos, celebrity cameos and musical guests, it always keeps us on our toes.

September 14, 2026 – 10:44 AM

Hello from San Francisco! We're excited for a week of Dreamforce, so stay tuned for all the latest updates as they happen.

The event kicks off tomorrow (Tuesday 15) so we'll try and get a sneak peek at the site (affectionally known as the Campgrounds) and anything else we spy.

• Revolut fell for a spoofed government email scam, leaking sensitive customer data to attackers
• Compromised info includes IDs, selfies, account statements, IBANs, and full transaction histories
• Criminals now allegedly leaking data on Telegram, demanding 10,000 BTC (~$780M) ransom from Revolut

Digital banking platform Revolut was tricked into giving away a treasure trove of sensitive customer data to hackers, and it is now coming back to bite it.

The company told TechCrunch that it recently fell victim to a “sophisticated external impersonation scam” in which the threat actor “utilized a legitimate government agency domain email to submit fraudulent requests for information”.

In other words, the attackers either broke into, or spoofed, an email address belonging to the police, tax authorities, or other government bodies with statutory powers to demand information, and used them to demand Revolut hand over sensitive customer data.

Demanding ransom

Cybernews reports that the compromised data includes customers’ birth dates, postal and email addresses, occupation, phone numbers, and copies of identity documents. TechCrunch added that verification selfies, account statements, and transaction histories may have also been compromised, together with IBANs, withdrawal records, complete transaction histories, and Bitcoin transactions.

Should these reports be confirmed, this will be a bonafide fiasco for Revolut.

"Upon detection, ⁠we immediately blocked the address and ​alerted the relevant government agency as ​well as enforcement agencies, data protection, and financial regulators," a company spokesperson told Reuters over the weekend.

So far, we don’t know exactly how many people are affected. Revolut said it is a “very limited” number, and that all of them had been notified already.

According to Coin Bureau, the criminals have started leaking sensitive data on Telegram, in a bid to pressure Revolut into paying a ransom demand. The publication shared screenshots of the threat actors apparently leaking a selfie and “full KYC” of a CEO of a crypto casino website, saying that the crooks are now demanding 10,000 BTC in exchange for deleting the data.

This would put the ransom demand at approximately $780 million which is obscene even by criminal standards.

"This one is deeply concerning and the implications for affected customers go well beyond a standard data breach notification," said Muhammad Yahya Patel, vCISO & Cybersecurity Advisor at Huntress. "Passports, driver's licenses, verification selfies, account statements, transaction histories, birth dates, addresses that's not a data leak, that's a complete identity theft kit handed to whoever sent those fraudulent requests."

"Every single component needed to impersonate someone, open accounts in their name, or bypass checks at other financial institutions is in that package. On the dark web, that kind of profile doesn't sell as individual records it sells as a ready-made fraud pack, and it commands a significant premium precisely because of its completeness."

"For a fintech built on digital identity verification, the bar for verifying third-party data requests should be exceptionally high. The question isn't why an attacker tried this. It's why a regulated financial institution handling highly sensitive data didn't have sufficiently rigorous verification controls to catch it," Patel concluded.

Via Cybernews

• Revolut fell for a spoofed government email scam, leaking sensitive customer data to attackers
• Compromised info includes IDs, selfies, account statements, IBANs, and full transaction histories
• Criminals now allegedly leaking data on Telegram, demanding 10,000 BTC (~$780M) ransom from Revolut

Digital banking platform Revolut was tricked into giving away a treasure trove of sensitive customer data to hackers, and it is now coming back to bite it.

The company told TechCrunch that it recently fell victim to a “sophisticated external impersonation scam” in which the threat actor “utilized a legitimate government agency domain email to submit fraudulent requests for information”.

In other words, the attackers either broke into, or spoofed, an email address belonging to the police, tax authorities, or other government bodies with statutory powers to demand information, and used them to demand Revolut hand over sensitive customer data.

Demanding ransom

Cybernews reports that the compromised data includes customers’ birth dates, postal and email addresses, occupation, phone numbers, and copies of identity documents. TechCrunch added that verification selfies, account statements, and transaction histories may have also been compromised, together with IBANs, withdrawal records, complete transaction histories, and Bitcoin transactions.

Should these reports be confirmed, this will be a bonafide fiasco for Revolut.

"Upon detection, ⁠we immediately blocked the address and ​alerted the relevant government agency as ​well as enforcement agencies, data protection, and financial regulators," a company spokesperson told Reuters over the weekend.

So far, we don’t know exactly how many people are affected. Revolut said it is a “very limited” number, and that all of them had been notified already.

According to Coin Bureau, the criminals have started leaking sensitive data on Telegram, in a bid to pressure Revolut into paying a ransom demand. The publication shared screenshots of the threat actors apparently leaking a selfie and “full KYC” of a CEO of a crypto casino website, saying that the crooks are now demanding 10,000 BTC in exchange for deleting the data.

This would put the ransom demand at approximately $780 million which is obscene even by criminal standards.

"This one is deeply concerning and the implications for affected customers go well beyond a standard data breach notification," said Muhammad Yahya Patel, vCISO & Cybersecurity Advisor at Huntress. "Passports, driver's licenses, verification selfies, account statements, transaction histories, birth dates, addresses that's not a data leak, that's a complete identity theft kit handed to whoever sent those fraudulent requests."

"Every single component needed to impersonate someone, open accounts in their name, or bypass checks at other financial institutions is in that package. On the dark web, that kind of profile doesn't sell as individual records it sells as a ready-made fraud pack, and it commands a significant premium precisely because of its completeness."

"For a fintech built on digital identity verification, the bar for verifying third-party data requests should be exceptionally high. The question isn't why an attacker tried this. It's why a regulated financial institution handling highly sensitive data didn't have sufficiently rigorous verification controls to catch it," Patel concluded.

Via Cybernews

• Highly exposed occupations are seeing less wage growth due to AI productivity gains
• 5.8 million US workers are considered to be working in these roles
• The impacts could lead to reduced living standards

While sceptics have been warning of mass layoffs for years, a new report reveals that AI could actually be leading to weaker wage growth instead of human replacement.

US wage growth is already slowing compared with inflation, and a new study by Apollo Global Management has proposed that AI could be to blame.

According to the data, real wage growth in highly exposed occupations has been around 6.7 percentage points lower than less-exposed occupations since 2023, while employment itself is largely no different.

What are the real implications of AI on jobs?

The white paper estimates that 5.8 million US workers, or around 3.7% of the US labor force, works in highly exposed jobs, and this is on the basis of earlier Anthropic work whereby at least half of the job's tasks can be replaced by Anthropic's AI tools.

"Companies are capturing AI productivity gains through wage compression rather than workforce reduction," Chief Economist Torsten Slok suggested. In other words, workers are getting more work done thanks to AI, so demand for additional labor has fallen and workers' bargaining powers have weakened.

Apollo estimates the effects of this could translate to around $28 billion in lost annual labor income, but this could be on the conservative side.

Crucially, Apollo's analysis shows that AI is more likely to be affecting the price of labor rather than the quantity of labor as human-AI hybrid workforces become more common.

Looking ahead, Apollo is warning that "profound implications for income inequality and living standards" could threaten workers globally.

While the research doesn't offer specific remedies, it does point toward government involvement and policymaking in order to slow down how quickly the transition happens and to offer greater support to impacted workers.

The Bambu Lab H2S is one of the largest and fastest enclosed 3D printers we've ever tested, and now it's got a very welcome discount in the Bambu lab 3D printer sale.

• Best for: Enthusiasts, engineers, and ecommerce businesses who need speedy 3D printers for larger- and higher-volume production.

• The deal: The Bambu Lab H2S AMS Combo 3D printer is down to $1399 (was $1500) at Bambu Lab. In the UK, the H2S is now £1099 (was £1199).

• Why it matters: We scored the Bambu Lab H2S 4.5 stars and awarded it a TechRadar Pro Recommends badge for easy setup, high speeds, and incredible accuracy. This CoreXY 3D printer has a large 340 x 320 x 340mm build volume, 350°C nozzle with 65°C active chamber heating, and 23-sensor AI-driven monitoring. In our final verdict, we found "the test prints that came out of this machine were some of the cleanest I've seen, and at this scale and consistency, it's really difficult to match with any other model."

• The catch: This deal is for the AMS Combo, unlocking multi-color printing out of the box. While the standalone 3D printer is also on sale for a cheaper price, you'll need to get the AMS accessory later if you want multi-color 3D printing.

H2S: was $1499.99 now $1399.99
340 x 320 x 340mm build volume, CoreXY mechanics rated up to 1000mm/s toolhead speed with 20,000mm/s² acceleration, a fully enclosed chamber with 65°C active heating and a 350°C nozzle for engineering materials, and 23 onboard sensors with 3 cameras for AI-driven print monitoring. Includes the printer, a textured PEI build plate, a spool holder, and an accessory box.

In the UK: now £1099 (was £1199)View Deal

Why we recommend it

Expert verdict

Our 3D printing expert Alastair was seriously impressed with the H2S during his 4.5 star review, saying "the H2S is impressive, and the single-nozzle hotend is highly tuned for precision and quality."

Benchmark quality

Accuracy is especially good, with Alastair noting: "One of the big points I would highlight while using this printer was not only the speed and scale, but also the quality of the surface finish. Surfaces were exceptionally smooth, and when I printed the Autodesk Kickstarter test model I've been using for many years, it was almost a full 30 out of 30."

Scale & speed

The build volume stands out. At 340 x 320 x 340mm, it is significantly larger than Bambu Lab's X1 Carbon, and the enclosed, heated chamber means it can handle engineering-grade filaments like ABS and nylon that need stable temperatures to print reliably. Combined with a toolhead speed up to 1000mm/s, this printer is built for people producing parts in volume, not just occasional hobby prints.

Features

The 23-sensor monitoring system is a meaningful reliability feature - the filament detection, for example, worked incredibly well during testing. This deal also includes the AMS Combo, letting you run multi-color models as soon as you're set up.

Price Context & Historical Value

What makes it worth flagging is that $1,399 for a printer with this build volume, this print speed, and a heated chamber remains a strong price relative to enclosed printers with comparable specs from other brands, several of which cost noticeably more for a smaller build area.

Should you buy it?

✅ Buy the Bambu Lab H2S AMS Combo if...

You want one of the largest build volumes available in an enclosed printer at this price, you print in engineering materials like ABS or nylon that need a heated chamber, or you need high-speed 3D printing for production.

❌ Skip the Bambu Lab H2S AMS Combo if...

You're a beginner or hobbyist (it's pricey for more casual users), or you're happy with single-color printing. If you don't need multi-color get the standalone model without the AMS Combo, as you can always upgrade later.

The Catch: What to know before you buy

This is a single-nozzle printer, so multicolor and multimaterial printing is slower and generates more waste than on Bambu Lab's dual-nozzle H2D models.

While setup and use is straightforward if you're familiar with 3D printers, there is a learning curve for anyone trying to push the printer's fastest custom speed settings, and it takes up significant desk space given its size.

This is a printer built for design studios, small businesses, and serious engineering hobbyists, not casual weekend printing.

• See all 3D printers at Bambu Lab

Over the last decade, digital identity has primarily been linked with the internet security industry.

Businesses have been focusing heavily on fraud prevention, profile authentication and compliance measures, all designed to answer a relatively simple question - is this person who they claim to be?

But as AI transforms how consumers discover, engage and purchase from brands, digital identity has become something much more integral for businesses across the board.

It has become a critical layer of business infrastructure, one that sits at the intersection of trust, experience, personalization and growth.

Ultimately, the businesses that understand this shift, will be better positioned to compete in an increasingly automated digital world.

And those that don’t, are likely to find themselves struggling to build trust, protect margins and maintain direct relationships with their customer base.

Proving Authenticity in a Complex AI Era

When we take a glance back, the internet was never designed with robust identity systems in mind. For decades, businesses relied heavily on a combination of usernames, passwords, email addresses and manual verification processes to establish trust online.

Whilst imperfect, and at times vastly time consuming, these processes were efficient in an era of human driven interactions. However, in the world of AI, this is no longer the case.

Artificial intelligence is dramatically lowering the barriers to creating fake accounts, synthetic identities and automated purchasing behavior. At the same time, consumers are becoming more conscious of how their personal information is collected, stored and used.

This creates a difficult balancing act for businesses. They need greater confidence that customers are genuine, while customers expect less friction and stronger privacy protections.

The traditional response has often been to collect more information. Yet in many cases, more data creates more risk. Every additional piece of personal information collected becomes another asset that must be protected, governed and justified.

As cyber threats increase and regulatory scrutiny grows, businesses are beginning to recognize that the future of identity may not be about collecting more data, but about collecting less.

The Future of Verification: Privacy by Design

One of the most important shifts happening within digital identity is the move away from document-heavy verification models towards trusted verification networks. Rather than asking users to repeatedly upload sensitive documents or provide excessive personal information, businesses can increasingly verify eligibility or identity through trusted institutional sources.

This approach delivers benefits for both organizations and consumers. For businesses, it reduces the operational and security burden associated with storing personally identifiable information. For consumers, it creates a faster, less intrusive experience. The key principle is simple: verify only what is necessary.

If a retailer needs to confirm that someone is a student, for example, they don't necessarily need access to every piece of information contained within a university record. They simply need confidence that the individual meets the criteria required to access a student offer. This is where audience verification platforms are becoming increasingly valuable.

In practice, this tends to happen in one of two ways. Sometimes verification works through a secure database lookup, where an individual's eligibility is checked against records already held by a relevant institution, such as an enrolment system, without the retailer ever seeing the underlying record itself, only a confirmation of status.

In other cases, it works through a trusted single-sign-on style login, where the individual authenticates directly with their own institution and that institution simply confirms their status back to the retailer, again without exposing any personal data to either party.

What both approaches have in common is that they lean on relationships of trust that already exist, rather than asking the consumer to prove who they are all over again.

The result is a more efficient exchange of trust between brands and consumers. In many ways, this reflects a broader evolution in digital identity. The goal is no longer to know everything about a user. The goal is to know enough to establish trust and ultimately, these notions of trust, are quickly becoming an ever-growing business growth strategy.

Trust has become a differentiator

Consumers are only becoming more demanding as they increasingly expect personalized experiences, relevant offers and seamless interactions. Yet, they also want transparency, control and confidence that their information is being handled responsibly. The brands that successfully balance these expectations gain more than compliance benefits. They build stronger customer relationships. Audience verification offers a useful example of this dynamic.

For a student, the value proposition is straightforward. They want access to relevant discounts and offers from brands they trust. The verification process itself is largely invisible, provided it is fast and frictionless. For the brand, however, the stakes are higher. Verification helps ensure promotional budgets reach the intended audience, protects against misuse and enables more effective customer acquisition.

Both sides benefit from a trusted exchange of value. This is why digital identity management should no longer be viewed purely through the lens of security teams. It has become a strategic capability that influences marketing effectiveness, customer acquisition, loyalty and brand trust.

Preparing for an AI-driven future

The rise of AI introduces another important dimension to the identity conversation. As large language models (LLMs) and AI assistants increasingly shape how consumers discover products and services, businesses face a new challenge: maintaining trusted, direct relationships with their audiences.

In the future, many interactions may occur through AI-powered intermediaries rather than traditional websites or apps. Discovery, recommendation and even purchasing decisions could increasingly be influenced by intelligent systems acting on behalf of consumers. In that environment, trusted audience data and verified customer relationships become even more valuable.

Through such technology, businesses can confidently identify and understand their audiences and will be better equipped to personalize experiences, deliver relevant recommendations and maintain authenticity across increasingly fragmented digital ecosystems.

Those that rely solely on broad marketing reach may find themselves losing visibility as AI-driven discovery mechanisms become more influential. The organizations that thrive will be those that treat digital identity not as a compliance requirement but as a foundational capability underpinning trust, personalization and growth.

From Identity Verification to Trust Infrastructure

The next chapter of digital commerce will be defined by trust. Not trust built through lengthy forms, excessive data collection or intrusive verification processes. Trust built through intelligent, privacy-conscious systems that enable businesses to verify what matters while respecting consumer expectations.

Digital identity is no longer simply about proving who someone is. It is about enabling secure, trusted and meaningful relationships between businesses and the people they serve. As AI continues to reshape commerce, organizations that invest in modern verification frameworks and data-minimization integrations will be better positioned to navigate the changes ahead.

Ultimately, the future belongs to businesses that can answer a simple but increasingly important question: how do you establish trust without creating friction? And clearly… digital identity is a strong bet.

We've listed the 8 best ecommerce platforms.

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

.com has long been established as the undisputed top dog when it comes to domains. Yet in 2026, a clear shift is underway in how brands select Top-Level Domains (TLDs) – with the likes of .ai, .io, and .store all seeing elevated rates of adoption.

These trends became transparent in Wix's study of domain registration trends in 2026. So, I spoke with Inbal Geva, Product Lead and GM at Wix Domains, to dive a little deeper into how TLD selection is starting to change and how businesses can make the most of this opportunity.

The study shows that .store websites drive 10% more sales than .com - why do you think that is?

One possibility is that this domain extension tells site visitors exactly what they are getting from the site (a store), which could mean more qualified shoppers are visiting the site. It also implies that the store owner is very clear about what they are selling on their site.

It also shows that .org websites are 34% more likely to drive ecommerce revenue than .com websites. What does the data tell us about why this might be?

The data shows an interesting correlation, but it doesn’t necessarily mean that the .org extension itself drives more ecommerce revenue.

.org is commonly associated with established organizations, nonprofits, and mission-driven businesses, which may have stronger communities and more engaged audiences. That could contribute to higher ecommerce activity.

So rather than attributing the difference to the extension itself, we see it as an indication that the types of organizations using .org — and the audiences they attract — may behave differently online.

.store, .online, and .shop registrations doubled in five months; does this highlight a shift in how businesses want to be represented online? Or is it more likely to reflect supply and affordability issues with .com domains?

So far, our data and the suggestion tests we’ve run consistently show that users still strongly prefer established TLDs, especially .com. So I wouldn’t conclude from the growth of .store, .online, and .shop that users now prefer changing the extension over adjusting the domain name.

What’s more interesting is that we’re seeing growing adoption of alternative TLDs, but .com remains the clear first choice. The growth of specific extensions can be driven by several factors — better availability of desirable names, pricing and promotions, increased visibility and distribution, and stronger marketing by the registries themselves.

Is .com still the default TLD? Or does it face a threat from others?

People clearly prefer a .com domain when their brand name is available.

Inbal Geva, Product Lead & GM at Wix Domains

.com is still the default.

People clearly prefer a .com domain when their brand name is available. Many users will still make a small adjustment to the domain name in order to keep .com, so I wouldn’t say it’s under threat. But alternative TLDs can still be appealing if your ideal .com is already registered. We also see people having fun with extensions like .yoga for yoga studios, .bar for a bar, and so on.

.ai increased by +20% from January to May 2026, and .io increased by a massive 32x in May 2026. Are these increases tied to a specific batch of business or product launches, or does it signal a more generic rise in these tech-associated TLDs?

We’ve seen more interest in tech and AI-driven extensions. Wix’s States of Website report showed that creators aren't treating their domains as an afterthought, but rather as a decision that is made early on in the process.

They're committing to them faster, which implies that they are taking choosing a domain name seriously. These TLDs immediately signal what the brand is about before you even get to the website.

How does the use of TLDs like .shop, .io, .ai, and .online impact a website's ability to perform well in SEO, AEO, and GEO?

Certain industries congregate around certain TLDs.

Tech uses .io, and AI founders use .ai. This means that the TLD can act as an additional contextual signal for LLMs that are looking for answers on niche topics. From an SEO point of view, this can also mean that users are more likely to click through to sites where the domain matches their intent. All of this is great for growth and visibility.

Wix analyzed almost a million domain registrations. What trend or statistic did you find most surprising?

Our data shows that .com alternatives can still perform well for businesses.

Inbal Geva, Product Lead & GM at Wix Domains

Historically, .com alternatives were sometimes seen as second best. But our data shows that .com alternatives can still perform well for businesses.

eCommerce sites with a .store extension earn 10% more revenue than sites with a .com extension. This doesn’t mean .com is bad—again, it’s still the default—but it means alternative extensions could be worth considering.

.com has long been established as the undisputed top dog when it comes to domains. Yet in 2026, a clear shift is underway in how brands select Top-Level Domains (TLDs) – with the likes of .ai, .io, and .store all seeing elevated rates of adoption.

These trends became transparent in Wix's study of domain registration trends in 2026. So, I spoke with Inbal Geva, Product Lead and GM at Wix Domains, to dive a little deeper into how TLD selection is starting to change and how businesses can make the most of this opportunity.

The study shows that .store websites drive 10% more sales than .com - why do you think that is?

One possibility is that this domain extension tells site visitors exactly what they are getting from the site (a store), which could mean more qualified shoppers are visiting the site. It also implies that the store owner is very clear about what they are selling on their site.

It also shows that .org websites are 34% more likely to drive ecommerce revenue than .com websites. What does the data tell us about why this might be?

The data shows an interesting correlation, but it doesn’t necessarily mean that the .org extension itself drives more ecommerce revenue.

.org is commonly associated with established organizations, nonprofits, and mission-driven businesses, which may have stronger communities and more engaged audiences. That could contribute to higher ecommerce activity.

So rather than attributing the difference to the extension itself, we see it as an indication that the types of organizations using .org — and the audiences they attract — may behave differently online.

.store, .online, and .shop registrations doubled in five months; does this highlight a shift in how businesses want to be represented online? Or is it more likely to reflect supply and affordability issues with .com domains?

So far, our data and the suggestion tests we’ve run consistently show that users still strongly prefer established TLDs, especially .com. So I wouldn’t conclude from the growth of .store, .online, and .shop that users now prefer changing the extension over adjusting the domain name.

What’s more interesting is that we’re seeing growing adoption of alternative TLDs, but .com remains the clear first choice. The growth of specific extensions can be driven by several factors — better availability of desirable names, pricing and promotions, increased visibility and distribution, and stronger marketing by the registries themselves.

Is .com still the default TLD? Or does it face a threat from others?

People clearly prefer a .com domain when their brand name is available.

Inbal Geva, Product Lead & GM at Wix Domains

.com is still the default.

People clearly prefer a .com domain when their brand name is available. Many users will still make a small adjustment to the domain name in order to keep .com, so I wouldn’t say it’s under threat. But alternative TLDs can still be appealing if your ideal .com is already registered. We also see people having fun with extensions like .yoga for yoga studios, .bar for a bar, and so on.

.ai increased by +20% from January to May 2026, and .io increased by a massive 32x in May 2026. Are these increases tied to a specific batch of business or product launches, or does it signal a more generic rise in these tech-associated TLDs?

We’ve seen more interest in tech and AI-driven extensions. Wix’s States of Website report showed that creators aren't treating their domains as an afterthought, but rather as a decision that is made early on in the process.

They're committing to them faster, which implies that they are taking choosing a domain name seriously. These TLDs immediately signal what the brand is about before you even get to the website.

How does the use of TLDs like .shop, .io, .ai, and .online impact a website's ability to perform well in SEO, AEO, and GEO?

Certain industries congregate around certain TLDs.

Tech uses .io, and AI founders use .ai. This means that the TLD can act as an additional contextual signal for LLMs that are looking for answers on niche topics. From an SEO point of view, this can also mean that users are more likely to click through to sites where the domain matches their intent. All of this is great for growth and visibility.

Wix analyzed almost a million domain registrations. What trend or statistic did you find most surprising?

Our data shows that .com alternatives can still perform well for businesses.

Inbal Geva, Product Lead & GM at Wix Domains

Historically, .com alternatives were sometimes seen as second best. But our data shows that .com alternatives can still perform well for businesses.

eCommerce sites with a .store extension earn 10% more revenue than sites with a .com extension. This doesn’t mean .com is bad—again, it’s still the default—but it means alternative extensions could be worth considering.

After a less-than-enthusiastic response to 2023's 360 Vis Nav, Dyson got the wheels rolling on its return to the robot vacuum market last year with its Spot+Scrub Ai. Now it seems momentum is building: this September the brand unveiled not one, but three new bots that would be joining the lineup.

There's the flagship R3 Nurovi Spot+Scrub UV — a successor to the Spot+Scrub Ai — alongside the R2 Nurovi Wash+Dry, and the R1 Nurovi Dry. I got a peek at all three at Dyson's IFA-adjacent launch event earlier in the month, and was intrigued to see some unique new features.

To get the inside scoop, I caught up with Nathan Lawson Mclean, a Dyson senior design manager who works across the brand's home products. Here are the five most surprising things I learned about the new Nurovi robot vacuum range...

1. Triangle mops have been in Dyson's invention cupboard for years

The R2 Nurovi Wash+Dry has a rather unusual mopping setup. Instead of two circular discs, it uses a pair of rounded 'Reuleaux' triangles. These rotate 'eccentrically' — that is, the point of rotation isn't central. The pads move like a rotary engine, or a Spirograph, depending on your preferred analogy. The benefit is the shape can get much closer into corners.

What's interesting is that this design has been in Dyson's back pocket for a while now, just waiting to be added to a suitable product.

"These have always been in development," says Lawson Mclean. "I remember the first Reuleaux pads — those were initially developed back in 2019. We put it on the front of a brush bar in a cleaner just to try and get into corners. And what it ended up doing, you can imagine, [it was] a bit like riding a bicycle with square wheels. It caused too much ricocheting."

As shown above, one of the mop pads can extend to the side to get closer to room edges (Image credit: Future)

This is apparently not an unusual scenario at Dyson. "We're actually always developing these technologies in tandem," he continues. "It's just about what's the right timing; the platform that we're able to integrate it on. You'll see that we're starting to get to the point where we're packing [new products] full of our unique Dyson technologies."

2. Not everyone wanted the side sweepers

All three of the new Nurovi robovacs have distinctive side sweepers. They feature a design that Lawson Mclean calls 'origami-inspired', but to me, looks more like a fish's fin. Apparently, the brand founder wasn't a fan originally.

"Side sweepers were always a bit of a faux pas with James Dyson," reveals Lawson Mclean. "He hated them. He thought they scattered dirt everywhere." For that reason, the brand's early models didn't have side sweepers at all — the 360 Vis Nav had a side duct that poked out when the bot sensed it was near the edge of a room.

"Side sweepers are fantastic. They are just a little chaotic. They also tend to wear out," continues Lawson Mclean. "More importantly, in James' eyes, they [end up getting] wrapped and covered in hair, which is a bit gross."

(Image credit: Future)

For the Nurovi, the Dyson engineers reimagined the traditional sweeper style to solve these issues, using "some of the advanced geometry" from the brand's conical cleaner heads found on the likes of the Dyson V16.

"It kind of has this really interesting over-molded pattern… that holds the nylon bristles at a very specific angle on the floor, helping us to avoid scattering this dust around, and corral things in a more controlled manner. It also stops hair wrapping around it so readily, too," says Lawson Mclean. "So it's a small, simple, effective, quality of life improvement."

3. UV will light up spills 'like a bowling alley'

The big, flashy upgrade found on the flagship R3 Nurovi Spot+Scrub UV is… well, UV. A UV light mounted onto the bot will highlight dried-on, invisible substances, ranging from sugary drinks to milk, tonic water and sparkling water. The R3 will then 'scrub' back and forth over the area to clear the mess before it gets sticky, or starts to smell.

"As the robot's running around your home on its mission, it will now expose these totally invisible stains that become breeding grounds for bacteria," explains Lawson Mclean. "You'd be surprised what fluoresces — a bit like a bowling alley in your own home.

4. The R1 is for those who don't trust robo-mops

It's increasingly rare to see a vacuum-only robot vac on the market these days, but Dyson has included one in its new lineup: the R1 Nurovi Dry. There were a multitude of reasons for this bot to exist, says Lawson Mclean. It's great for people who don't have have hard flooring in their home, or are looking for a more affordable option, or simply want to avoid — as he puts it — "a great big bloody dock like some robots have".

But it's also designed as an option for those who are wary of robo-mops. "There are many people who — I know for a fact, through user trials — don't trust robotic vacuums with their carpets. You know… we've ruined many of James' Persian rugs over the years," he recalls, presumably referring to brand founder James Dyson. "[It's] perhaps even a bit of an entry gateway to robotic vacuums, converting those people that don't really see the value of wet and dry robots just yet."

The Dyson R1 Nurovi (left) is a vacuum-only sister model to the wet-and-dry R2 (right) (Image credit: Future)5. One major update has nothing to do with the design

There are plenty of interesting design features, but it might turn out that one of the most significant upgrades is nothing to do with the robot itself at all. "I think when we talk about, 'Why are people not using this fantastic technology?', there's a lot of barriers. There are, just really pragmatically," says Lawson Mclean. "Even, I think, with people [in their 30s], they go, yeah, I don't really understand it, I don't really know what the setup is like."

Dyson's response has been to set up a 'white glove' onboarding service, which takes the form of an actual human who will, via video call, walk you through the setup process. That includes everything from unboxing and connecting the bot to mapping your home properly.

It's part of the new 'Dyson Concierge' service, and it's free with your purchase. (At time of writing, the service is being advertised on Dyson US, but not yet on the UK or Australian sites — I'm confirming if and when it'll be available in these territories.)

"When we started doing it, we just saw this huge uplift," recalls Lawson Mclean. "It wasn't a problem with the robots. It was just a problem with people having poor maps to begin with because perhaps they left a door closed, you know?"

Take a look at the full Dyson robovac range

Below you can see the Dyson robot vacuums currently available in your country.

Note: the R3 Spot+Scrub UV is not yet available to buy. In the US, it was meant to launch exclusively at CostCo in early September, but at time of writing (14 Sept) is hasn't arrived yet. It's due to arrive in the UK and Australia in 'early 2027'.

Dyson R1 Nurovi Dry

Dyson R2 Nurovi Wash+Dry

Dyson Spot+Scrub Ai Robot Vacuum

The Spot+Scrub Ai Robot Vacuum is Dyson's previous-gen robot vacuum — it'll be superseded by the R3 Nurovi Spot+Scrub UV when that model becomes available. It has a similar feature where it'll 'spot' stains and 'scrub' back and forth over them, but it's not powered by UV light.

Read our full review

Pros

• Good hardware
• Great at little-and-often cleaning
• Excellent self-cleaning
• Smart obstacle avoidance
• Auto vacuuming works well
• Bagless design

Cons

• Some navigation issues
• Unreliable self-empty
• Loud in every mode
• Large vacuum and dock
• Hair wrap on roller
• Transparent bins display dirt

After a less-than-enthusiastic response to 2023's 360 Vis Nav, Dyson got the wheels rolling on its return to the robot vacuum market last year with its Spot+Scrub Ai. Now it seems momentum is building: this September the brand unveiled not one, but three new bots that would be joining the lineup.

There's the flagship R3 Nurovi Spot+Scrub UV — a successor to the Spot+Scrub Ai — alongside the R2 Nurovi Wash+Dry, and the R1 Nurovi Dry. I got a peek at all three at Dyson's IFA-adjacent launch event earlier in the month, and was intrigued to see some unique new features.

To get the inside scoop, I caught up with Nathan Lawson Mclean, a Dyson senior design manager who works across the brand's home products. Here are the five most surprising things I learned about the new Nurovi robot vacuum range...

1. Triangle mops have been in Dyson's invention cupboard for years

The R2 Nurovi Wash+Dry has a rather unusual mopping setup. Instead of two circular discs, it uses a pair of rounded 'Reuleaux' triangles. These rotate 'eccentrically' — that is, the point of rotation isn't central. The pads move like a rotary engine, or a Spirograph, depending on your preferred analogy. The benefit is the shape can get much closer into corners.

What's interesting is that this design has been in Dyson's back pocket for a while now, just waiting to be added to a suitable product.

"These have always been in development," says Lawson Mclean. "I remember the first Reuleaux pads — those were initially developed back in 2019. We put it on the front of a brush bar in a cleaner just to try and get into corners. And what it ended up doing, you can imagine, [it was] a bit like riding a bicycle with square wheels. It caused too much ricocheting."

As shown above, one of the mop pads can extend to the side to get closer to room edges (Image credit: Future)

This is apparently not an unusual scenario at Dyson. "We're actually always developing these technologies in tandem," he continues. "It's just about what's the right timing; the platform that we're able to integrate it on. You'll see that we're starting to get to the point where we're packing [new products] full of our unique Dyson technologies."

2. Not everyone wanted the side sweepers

All three of the new Nurovi robovacs have distinctive side sweepers. They feature a design that Lawson Mclean calls 'origami-inspired', but to me, looks more like a fish's fin. Apparently, the brand founder wasn't a fan originally.

"Side sweepers were always a bit of a faux pas with James Dyson," reveals Lawson Mclean. "He hated them. He thought they scattered dirt everywhere." For that reason, the brand's early models didn't have side sweepers at all — the 360 Vis Nav had a side duct that poked out when the bot sensed it was near the edge of a room.

"Side sweepers are fantastic. They are just a little chaotic. They also tend to wear out," continues Lawson Mclean. "More importantly, in James' eyes, they [end up getting] wrapped and covered in hair, which is a bit gross."

(Image credit: Future)

For the Nurovi, the Dyson engineers reimagined the traditional sweeper style to solve these issues, using "some of the advanced geometry" from the brand's conical cleaner heads found on the likes of the Dyson V16.

"It kind of has this really interesting over-molded pattern… that holds the nylon bristles at a very specific angle on the floor, helping us to avoid scattering this dust around, and corral things in a more controlled manner. It also stops hair wrapping around it so readily, too," says Lawson Mclean. "So it's a small, simple, effective, quality of life improvement."

3. UV will light up spills 'like a bowling alley'

The big, flashy upgrade found on the flagship R3 Nurovi Spot+Scrub UV is… well, UV. A UV light mounted onto the bot will highlight dried-on, invisible substances, ranging from sugary drinks to milk, tonic water and sparkling water. The R3 will then 'scrub' back and forth over the area to clear the mess before it gets sticky, or starts to smell.

"As the robot's running around your home on its mission, it will now expose these totally invisible stains that become breeding grounds for bacteria," explains Lawson Mclean. "You'd be surprised what fluoresces — a bit like a bowling alley in your own home.

4. The R1 is for those who don't trust robo-mops

It's increasingly rare to see a vacuum-only robot vac on the market these days, but Dyson has included one in its new lineup: the R1 Nurovi Dry. There were a multitude of reasons for this bot to exist, says Lawson Mclean. It's great for people who don't have have hard flooring in their home, or are looking for a more affordable option, or simply want to avoid — as he puts it — "a great big bloody dock like some robots have".

But it's also designed as an option for those who are wary of robo-mops. "There are many people who — I know for a fact, through user trials — don't trust robotic vacuums with their carpets. You know… we've ruined many of James' Persian rugs over the years," he recalls, presumably referring to brand founder James Dyson. "[It's] perhaps even a bit of an entry gateway to robotic vacuums, converting those people that don't really see the value of wet and dry robots just yet."

The Dyson R1 Nurovi (left) is a vacuum-only sister model to the wet-and-dry R2 (right) (Image credit: Future)5. One major update has nothing to do with the design

There are plenty of interesting design features, but it might turn out that one of the most significant upgrades is nothing to do with the robot itself at all. "I think when we talk about, 'Why are people not using this fantastic technology?', there's a lot of barriers. There are, just really pragmatically," says Lawson Mclean. "Even, I think, with people [in their 30s], they go, yeah, I don't really understand it, I don't really know what the setup is like."

Dyson's response has been to set up a 'white glove' onboarding service, which takes the form of an actual human who will, via video call, walk you through the setup process. That includes everything from unboxing and connecting the bot to mapping your home properly.

It's part of the new 'Dyson Concierge' service, and it's free with your purchase. (At time of writing, the service is being advertised on Dyson US, but not yet on the UK or Australian sites — I'm confirming if and when it'll be available in these territories.)

"When we started doing it, we just saw this huge uplift," recalls Lawson Mclean. "It wasn't a problem with the robots. It was just a problem with people having poor maps to begin with because perhaps they left a door closed, you know?"

Take a look at the full Dyson robovac range

Below you can see the Dyson robot vacuums currently available in your country.

Note: the R3 Spot+Scrub UV is not yet available to buy. In the US, it was meant to launch exclusively at CostCo in early September, but at time of writing (14 Sept) is hasn't arrived yet. It's due to arrive in the UK and Australia in 'early 2027'.

Dyson R1 Nurovi Dry

Dyson R2 Nurovi Wash+Dry

Dyson Spot+Scrub Ai Robot Vacuum

The Spot+Scrub Ai Robot Vacuum is Dyson's previous-gen robot vacuum — it'll be superseded by the R3 Nurovi Spot+Scrub UV when that model becomes available. It has a similar feature where it'll 'spot' stains and 'scrub' back and forth over them, but it's not powered by UV light.

Read our full review

Pros

• Good hardware
• Great at little-and-often cleaning
• Excellent self-cleaning
• Smart obstacle avoidance
• Auto vacuuming works well
• Bagless design

Cons

• Some navigation issues
• Unreliable self-empty
• Loud in every mode
• Large vacuum and dock
• Hair wrap on roller
• Transparent bins display dirt

Corsair's Vengeance RBG memory has consistently impressed us in use, and with the on-going memory crisis sending prices sky-rocketing, we're constantly on the look-out for discounts on its speedy DDR5 RAM. And we've found on on the the white 32GB Corsair Vengeance RGB DDR5-6000 kit over at Newegg.

• Best for: PC builders and gamers who want 32GB of fast DDR5 memory for a compatible Intel system.

• The deal: The white Corsair Vengeance RGB 32GB DDR5-6000 kit is $480 (was $500) at Newegg when you use code GGUF239

• Why it matters: You get two 16GB DDR5-6000 modules with CL36-44-44-96 timings, Intel XMP 3.0 support and onboard voltage regulation. Each module also has ten individually addressable RGB LEDs that you can customize through Corsair iCUE.

• The catch: You'll need a DDR5-compatible motherboard, with Corsair recommending Intel 600 and 700 series models. You'll also typically need to enable XMP in the BIOS to reach the advertised speed.

32GB Vengeance RGB DDR5-6000 RAM: was $499.99 now $479.99
This 32GB dual-channel kit includes two 16GB DDR5-6000 modules with CL36-44-44-96 timings, 1.40V operation, Intel XMP 3.0, aluminum heat spreaders, ten-zone RGB lighting, iCUE support and onboard voltage regulation.

Use coupon GGUF239 to get the cheapest price.View Deal

Also consider

32GB Vengeance RGB DDR5-6000 RAM: was $599.68 now $490.87
This 32GB dual-channel kit includes two 16GB DDR5-6000 modules with CL36 timings, Intel XMP 3.0 support, onboard voltage regulation and ten individually addressable RGB LEDs per module for customizable lighting.View Deal

Why we recommend it

A 32GB dual-channel kit provides plenty of memory for a modern PC, giving you room for demanding multitasking, content creation and other memory-intensive applications. The DDR5-6000 speed provides a good balance between bandwidth and compatibility, and Intel XMP 3.0 support makes it easier to run the memory at its advertised settings without manually tweaking values in the BIOS.

Price context & historical value

Newegg currently lists the 32GB kit at $499.99, with promo code GGUF239 reducing the checkout price by $20 to $479.99. The current $479.99 price isn't the lowest we've seen for this exact kit however. It was $439.99 last autumn.

RAM prices can vary considerably right now, so the value here depends heavily on wanting this particular white RGB configuration rather than simply looking for the cheapest 32GB DDR5 kit.

Should you buy it?

✅ Buy the Corsair Vengeance RGB if...

You're building or upgrading a DDR5 Intel PC and want 32GB of fast memory that also looks good. DDR5-6000, XMP 3.0 and customizable RGB lighting make it particularly appealing for a build with a windowed case.

❌ Skip the Corsair Vengeance RGB if...

You're primarily interested in getting 32GB of DDR5 memory for the lowest possible price. You're paying extra for the white heat spreaders, RGB lighting and Corsair ecosystem.

The Catch: What to know before you buy

This is DDR5 memory, so it won't work in a motherboard that only accepts DDR4. Corsair specifically recommends the kit for compatible Intel platforms, including Intel 600 and 700 series motherboards.

The advertised DDR5-6000 specification also relies on an XMP performance profile, so won't necessarily run at that speed straight after installation. Its CL36-44-44-96 timings operate at 1.40V, so you'll need to enable the appropriate XMP profile in your BIOS to get the best settings.

• Shop all RAM at Newegg

Corsair's Vengeance RBG memory has consistently impressed us in use, and with the on-going memory crisis sending prices sky-rocketing, we're constantly on the look-out for discounts on its speedy DDR5 RAM. And we've found on on the the white 32GB Corsair Vengeance RGB DDR5-6000 kit over at Newegg.

• Best for: PC builders and gamers who want 32GB of fast DDR5 memory for a compatible Intel system.

• The deal: The white Corsair Vengeance RGB 32GB DDR5-6000 kit is $480 (was $500) at Newegg when you use code GGUF239

• Why it matters: You get two 16GB DDR5-6000 modules with CL36-44-44-96 timings, Intel XMP 3.0 support and onboard voltage regulation. Each module also has ten individually addressable RGB LEDs that you can customize through Corsair iCUE.

• The catch: You'll need a DDR5-compatible motherboard, with Corsair recommending Intel 600 and 700 series models. You'll also typically need to enable XMP in the BIOS to reach the advertised speed.

32GB Vengeance RGB DDR5-6000 RAM: was $499.99 now $479.99
This 32GB dual-channel kit includes two 16GB DDR5-6000 modules with CL36-44-44-96 timings, 1.40V operation, Intel XMP 3.0, aluminum heat spreaders, ten-zone RGB lighting, iCUE support and onboard voltage regulation.

Use coupon GGUF239 to get the cheapest price.View Deal

Also consider

32GB Vengeance RGB DDR5-6000 RAM: was $599.68 now $490.87
This 32GB dual-channel kit includes two 16GB DDR5-6000 modules with CL36 timings, Intel XMP 3.0 support, onboard voltage regulation and ten individually addressable RGB LEDs per module for customizable lighting.View Deal

Why we recommend it

A 32GB dual-channel kit provides plenty of memory for a modern PC, giving you room for demanding multitasking, content creation and other memory-intensive applications. The DDR5-6000 speed provides a good balance between bandwidth and compatibility, and Intel XMP 3.0 support makes it easier to run the memory at its advertised settings without manually tweaking values in the BIOS.

Price context & historical value

Newegg currently lists the 32GB kit at $499.99, with promo code GGUF239 reducing the checkout price by $20 to $479.99. The current $479.99 price isn't the lowest we've seen for this exact kit however. It was $439.99 last autumn.

RAM prices can vary considerably right now, so the value here depends heavily on wanting this particular white RGB configuration rather than simply looking for the cheapest 32GB DDR5 kit.

Should you buy it?

✅ Buy the Corsair Vengeance RGB if...

You're building or upgrading a DDR5 Intel PC and want 32GB of fast memory that also looks good. DDR5-6000, XMP 3.0 and customizable RGB lighting make it particularly appealing for a build with a windowed case.

❌ Skip the Corsair Vengeance RGB if...

You're primarily interested in getting 32GB of DDR5 memory for the lowest possible price. You're paying extra for the white heat spreaders, RGB lighting and Corsair ecosystem.

The Catch: What to know before you buy

This is DDR5 memory, so it won't work in a motherboard that only accepts DDR4. Corsair specifically recommends the kit for compatible Intel platforms, including Intel 600 and 700 series motherboards.

The advertised DDR5-6000 specification also relies on an XMP performance profile, so won't necessarily run at that speed straight after installation. Its CL36-44-44-96 timings operate at 1.40V, so you'll need to enable the appropriate XMP profile in your BIOS to get the best settings.

• Shop all RAM at Newegg

• Krafton announced PUBG: Dednet at Gamescom this year
• The '90s-inspired shooter features some bonkers abilities
• "We're kind of finding our own way here," explains game director Dave Curd

Players slithering along the ground like snakes, swords for hands, a tiny doll wielding a power drill driving in a toy car — some of these wild sights seen in the debut trailer for PUBG: Dednet seem completely at odds with the realistic and reserved approach the defining battle royale series is known for. When I bring this up to the game's creative director, Dave Curd, he tells me that is entirely the point.

"If you want to excite and delight people, you have to give them something fresh," he says me in an interview at Gamescom. "If you want a cool, stoic, mil-sim, take your pick; they're everywhere. But what everyone else can't do, and maybe they don't want to, is my take on this world, this lore, these characters, these abilities."

Curd explains that the tone of PUBG: Dednet is like "my version of the 90s filtered through a meatgrinder," and you can certainly see that from the debut trailer. It's zany, over-the-top, and drenched in color, with gunfights taking place in arcades, car chases through rural towns lit up by neon signs, and a grungy, punk vibe throughout. Safe to say, it stands out.

"The audience will forgive almost anything but being bored," suggests Curd. "So, I think this trailer will get a lot of love and hate.

"As long as they're feeling something passionately, I think we've done a good job," he continues. "But if people look past that and their eyes glaze over, then that would be worrying."

While Dednet is still effectively a battle royale, pitting multiple teams of three against each other in a contest to be the last ones standing, it continues to mix things up not just in the tone and character abilities, but also in how your game progresses over time with several roguelite elements.

It starts with the concept of runs, where you might nominate how many wins or placements you can achieve before you hit a certain number of losses. Win or lose, at the end of a match, you're presented with choices that can alter or upgrade your chosen character's abilities and play style.

Perhaps you'll lean into more of a support role to assist your teammates with heals or protection with a riot shield made out of a car door? Or maybe you'll take that aforementioned doll, who can ride on your back into battle or stealthily sit on your opponent's shoulder and be a nuisance.

(Image credit: Krafton)

"I'm pretty confident this isn't going to be in this year's Battlefield update," jokes Curd, when explaining how they came up with abilities like this for the multiplayer shooter. "I think our internal test is: does it make us laugh? Does it sound like an interesting challenge? And do we think about it the next day? We're kind of finding our own way here."

That neatly summarises how Dave Curd and the team are experimenting for PUBG: Dednet to create something unique and interesting in the space. It definitely caught my eye during this year's Gamescom, and if it got your attention too, you can pre-register for the closed beta now.

• Krafton announced PUBG: Dednet at Gamescom this year
• The '90s-inspired shooter features some bonkers abilities
• "We're kind of finding our own way here," explains game director Dave Curd

Players slithering along the ground like snakes, swords for hands, a tiny doll wielding a power drill driving in a toy car — some of these wild sights seen in the debut trailer for PUBG: Dednet seem completely at odds with the realistic and reserved approach the defining battle royale series is known for. When I bring this up to the game's creative director, Dave Curd, he tells me that is entirely the point.

"If you want to excite and delight people, you have to give them something fresh," he says me in an interview at Gamescom. "If you want a cool, stoic, mil-sim, take your pick; they're everywhere. But what everyone else can't do, and maybe they don't want to, is my take on this world, this lore, these characters, these abilities."

Curd explains that the tone of PUBG: Dednet is like "my version of the 90s filtered through a meatgrinder," and you can certainly see that from the debut trailer. It's zany, over-the-top, and drenched in color, with gunfights taking place in arcades, car chases through rural towns lit up by neon signs, and a grungy, punk vibe throughout. Safe to say, it stands out.

"The audience will forgive almost anything but being bored," suggests Curd. "So, I think this trailer will get a lot of love and hate.

"As long as they're feeling something passionately, I think we've done a good job," he continues. "But if people look past that and their eyes glaze over, then that would be worrying."

While Dednet is still effectively a battle royale, pitting multiple teams of three against each other in a contest to be the last ones standing, it continues to mix things up not just in the tone and character abilities, but also in how your game progresses over time with several roguelite elements.

It starts with the concept of runs, where you might nominate how many wins or placements you can achieve before you hit a certain number of losses. Win or lose, at the end of a match, you're presented with choices that can alter or upgrade your chosen character's abilities and play style.

Perhaps you'll lean into more of a support role to assist your teammates with heals or protection with a riot shield made out of a car door? Or maybe you'll take that aforementioned doll, who can ride on your back into battle or stealthily sit on your opponent's shoulder and be a nuisance.

(Image credit: Krafton)

"I'm pretty confident this isn't going to be in this year's Battlefield update," jokes Curd, when explaining how they came up with abilities like this for the multiplayer shooter. "I think our internal test is: does it make us laugh? Does it sound like an interesting challenge? And do we think about it the next day? We're kind of finding our own way here."

That neatly summarises how Dave Curd and the team are experimenting for PUBG: Dednet to create something unique and interesting in the space. It definitely caught my eye during this year's Gamescom, and if it got your attention too, you can pre-register for the closed beta now.

Today's the day when Apple starts rolling out iOS 27 to millions of iPhones around the world — and as I've been using the developer version of the software since it was first pushed out in June, I can tell you about the best features coming your way.

The good news is that every iPhone that can run iOS 26 can also run iOS 27, though some of the more advanced Siri AI features (like the ability to change its pace and 'expressivity') are only available on the newer models with more powerful chipsets.

You'll get a notification popping up on your phone when iOS 27 is ready for you, but if you're not patient enough for that you can open up Settings on iOS and tap General > Software Update to see if there's an update waiting.

1. Reframe your photos

(Image credit: Future / Apple)

iOS 27 arrives with some clever AI-powered photo editing tricks, including one called Reframe. It lets you slightly change the angle of a picture you've taken, with AI adjusting the background to match.

In Photos, tap on an image to open it, then tap the icon showing three sliders at the bottom. If you then tap Tools and Reframe, you can touch and drag on the screen to change the perspective (use two fingers to pan and zoom). Tap Reframe to confirm.

The other options on the Tools menu can be useful too: Extend will use AI to imagine a wider background canvas for you, while Clean Up lets you remove specific objects or people from your images.

2. Create Shortcuts with natural language

Creating shortcuts with natural language (Image credit: Future)

The Shortcuts app is one of the most powerful tools available in iOS, but as we've written about before, it's not all that easy to use. With the new software update, building shortcuts becomes much easier: you can just describe the combinations you want using natural language, which significantly simplifies the process.

Shortcuts let you set up a myriad of automations and routines on your iPhone: turn text into audio, play a podcast after your morning alarm, tag notes in batches, let people know you're running late, create GIFs from videos, and a whole lot more. Open the Gallery tab in the shortcuts app for some inspiration.

With iOS 27 installed, if you tap the + (plus) button on the main All Shortcuts screen, you can simply describe what you want your shortcut to do — whether it's attaching screenshots to emails or summarizing your to do list for the day.

3. Choose a separate alarm volume

Not a feature that's going to necessarily grab the headlines, but definitely one of the most useful ones that I've come across since I first installed iOS 27: you can now set alarm volumes separately from the main ringtone and system volumes.

Head into Sounds & Haptics from Settings on iOS, turn off the Match Ringtone Volume toggle switch under Alarms and Timers, and set your preferred alarm volume — a customization that's been long overdue.

4. Adjust the Liquid Glass effect

Adjusting Liquid Glass in iOS 27 (Image credit: Future)

The people asked, and Apple listened: you can now take the edge off the Liquid Glass effect if you find it a bit overwhelming in its default state. The available adjustments aren't huge, but you can tone it down (or turn it up) a little.

From Settings in iOS, tap Appearance > Liquid Glass. You get a slider that lets you shift from quite a lot of transparency (on the left) to hardly any transparency (on the right), with the changes previewed for you on the image above.

Apple has also done some cleaning up of Liquid Glass in iOS 27, so you should notice a few refinements as you start using the new operating system: as per Apple, Liquid Glass now offers "more uniform refraction and improved contrast", while individual app icons have been made "sharper and more detailed".

5. Siri AI

(Image credit: Future)

The biggest new feature in iOS 27 is the comprehensive upgrade that Siri's getting. It has a new name (the same as the old name, with "AI" on the end), plus vastly better underlying models (courtesy of Google) that mean it's now on a par with the likes of ChatGPT when it comes to the quality and breadth of answers you get.

So what's actually new? When it first lands, Siri AI will spend a while indexing everything on your phone, so it can give you more context about what's in your texts, calendars, and emails (none of this data gets sent back to Apple) — and it can take more actions inside apps too, including creating notes, rotating photos, resuming podcasts, and much more.

You can now give commands like "show me photos from when I went to Spain" with this personal context, and ask questions like "why do stars twinkle in the sky?" without the assistant handing off to ChatGPT. There's also a standalone Siri app for the first time, meaning you can save your chats and sync them across devices.

• Anthropic and OpenAI are asking governments to work together to slow risky innovation
• US politicians are factoring in AI into their arguments as midterm elections approach
• Trump currently worries that taking the foot off the throttle could let China overtake

Tech leaders at the forefront of AI, including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, are publicly calling for slow AI innovation in order to buy more time to assess AI risks.

Among the risks highlighted in Amodei's warning are that AI systems could become impossible for humans to control, cyberattacks could become worse and AI could even facilitate biological weaponry.

Subsequently, he's called for external evaluators to gain deeper access to their models and safety processes, but crucially, Amodei advocate for greater government involvement.

Even AI companies want governments to step in

Anthropic's CEO is pushing for democratic countries to agree common AI safety standards and rules, and the eventual international coordination over the most dangerous AI capabilities, including with China.

And it's this international coordination that could prove key to slowing down risky innovation, because to date, US President Donald Trump has pushed back against warnings over worries that slowing down American companies would ultimately cause China to overtake the US.

"Whoever wins AI wins," Trump said (via CNBC).

"If Congress just races in and does some sort of emergency session to try to regulate AI, we will lose the race to China," House Speaker Mike Johnson, R-La., said.

And for the US in particular, now is a more important time than ever. Democratic nominee for Governor of Texas Gina Hinojosa recently described data centers, and therefore AI, as "the fight of this election season."

There's an increasingly narrow pre-election window for legislation as the technology increasingly shows divides not just between the two big parties, but also within them.

"We've reached a new chapter in AI capabilities, and that demands a new chapter for AI policy," OpenAI Chief Global Affairs Officer Chris Lehane wrote. And with the likes of OpenAI and Anthropic, right at the top of American AI, warning about AI's potential impacts, Senator Ruben Gallego, D-Ariz., likened the scenario to Dr. Frankenstein "telling us the monster is escaping."

With midterm elections just two months away, we're likely to see discussions around AI ramp up as politicians and influential leaders continue to get involved in guardrail discourse.