News
- 30% of Britons are providing AI chatbots with confidential personal information
- Research from NymVPN shows company and customer data is also at risk
- Emphasizes the importance of taking precautions, like using a quality VPN
Almost one in three Britons shares sensitive personal data with AI chatbots like OpenAI’s ChatGPT, according to research from cybersecurity company NymVPN. 30% of Brits have fed AI chatbots with confidential information such as health and banking data, potentially putting their privacy – and that of others – at risk.
This oversharing with the likes of ChatGPT and Google Gemini comes despite 48% of respondents expressing privacy concerns over AI chatbots. This signals that the issue extends to the workplace, with employees sharing sensitive company and customer data.
NymVPN’s findings come in the wake of a number of recent high-profile data breaches, most notably the Marks & Spencer cyber attack, which shows just how easily confidential data can fall into the wrong hands.
“Convenience is being prioritized over security”NymVPN’s research reveals that 26% of respondents admitted to disclosing financial information related to salary, investments, and mortgages to AI chatbots. Riskier still, 18% shared credit card or bank account data.
24% of those surveyed by NymVPN admit to having shared customer data – including names and email addresses – with AI chatbots. More worrying still, 16% uploaded company financial data and internal documents such as contracts. This is despite 43% expressing worry about sensitive company data being leaked by AI tools.
“AI tools have rapidly become part of how people work, but we’re seeing a worrying trend where convenience is being prioritized over security,” said Harry Halpin, CEO of NymVPN.
M&S, Co-op, and Adidas have all been in the headlines for the wrong reasons, having fallen victim to data breaches. “High-profile breaches show how vulnerable even major organizations can be, and the more personal and corporate data that is fed into AI, the bigger the target becomes for cybercriminals,” said Halpin.
The importance of not oversharingSince nearly a quarter of respondents share customer data with AI chatbots, this emphasizes the urgency of companies implementing clear guidelines and formal policies for the use of AI in the workplace.
“Employees and businesses urgently need to think about how they’re protecting both personal privacy and company data when using AI tools,” said Halpin.
Although avoiding AI chatbots entirely would be the optimal solution for privacy, it’s not always the most practical. Users should, at the very least, avoid sharing sensitive information with AI chatbots. Privacy settings can also be tweaked, such as disabling chat history or opting out of model training.
A VPN can add a layer of privacy when using AI chatbots such as ChatGPT, encrypting a user’s internet traffic and original IP address. This helps keep a user’s location private and prevents their ISP from seeing what they’re doing online. Still, even the best VPN isn’t enough if sensitive personal data is still being fed to AI.
Agents – software systems capable of decision making or performing tasks autonomously - are no longer experimental. Today, these agents are operational, distributed and actively making decisions across the enterprise. From writing code to scheduling tasks, agents are starting to permeate every facet of business. The reason is clear: agents promise significant productivity gains.
Some will be deeply embedded, making them difficult to detect or monitor. Others will operate autonomously, continuously learning and adapting in real time. Many may have broad access privileges in the name of efficiency. This introduces significant potential for both positive impact and risk.
And as adoption grows, many organizations will face a new challenge: securing agents at scale. Businesses will need to ensure that innovation doesn’t outpace security and governance. The stakes are too high; one single misalignment, vulnerability or unintended behavior can lead to a runaway effect of unethical or harmful actions.
We’ve already seen real-world examples of AI failures – sometimes exposing sensitive data or making critical errors. One AI assistant notoriously advised users to eat rocks, and in another case a customer service chatbot deployed by a logistics company began issuing aggressive responses. Both examples show the risk of poor training data – AI agents don’t just learn facts, they learn behaviors, and bad input leads to bad output.
Cloud Déjà vu, Now with AgentsWithout consistent oversight, agents can act outside their intended use and damage brand reputation. That’s why it’s important that security is baked in at the start. Like salt and pepper, you can always sprinkle more on later, but if you forget to add it while cooking, the flavor – and in this case the protection – just won’t be the same. Security must be integrated from the outset, waiting until after deployment to retrofit security is a recipe for vulnerabilities.
Just consider what happened during the mass migration to cloud computing technologies. Adoption led to serious security missteps, data silos and visibility gaps. Gaps that have been and continue to be exploited by attackers today.
Now with agents it’s like a bad case of déjà vu. Once again, innovation is outpacing security. In many cases, these autonomous tools are being integrated into critical systems with limited oversight and lacking proper security and controls.
If we don’t apply the hard lessons learned from the cloud era we risk repeating the same mistakes, but this time with far more unpredictable systems. That’s why security must be at the core of agents.
Securing Every Agent TouchpointBut securing agents requires an expanded approach, one that accounts for autonomous behaviors, including those ongoing interactions with data, systems and users. Agents need a strong trust layer, where every interaction, from API calls to sensitive data handling, must be mapped, protected and governed in real time.
A core part of this trust layer is securing the data agents interact with—inputs, outputs and everything in between. Data is the fuel of agents, and without foundational security that fuel becomes a major risk. Enterprises must focus on the fundamentals like data discovery and classification, encryption and key management.
Access and Identity Management strategies must also evolve as agents take on more advanced roles in the enterprise. Like humans, every agent will require its own unique credentials, roles and permissions to ensure that every interaction is authorized and verified.
Agent credentials should be stored in a secure, automated credential vault, with policies enforcing regular rotation, access logging, and immediate revocation if misuse is detected. Organizations must be able to distinguish between agents using managed or unmanaged credentials.
And once agent credentials are brought under management, it's crucial to protect and enforce proper lifecycle management and governance. By provisioning, rotating, auditing, protecting and decommissioning credentials organizations can reduce the risk of credential misuse and theft.
Without strong identity oversight, businesses risk losing visibility of both human and agent identities and control over autonomous actions.
Decentralized Agents Need Centralized SecurityHowever, at scale, managing agents and especially autonomous ones, will require additional control to monitor behavior, interactions and deviations from policy. Consider a type of agent “security manager” that brings agents and humans on the loop to build trust in how agents operate.
This should be more than a dashboard, rather intelligence capable of understanding what agents are doing, why they’re doing it and whether their behavior aligns with policies and risk thresholds on a constant basis. It flags anomalies, enforces constraints and enables human review, when needed.
That last part is particularly important. Human oversight remains essential, especially when scaling agents. This control layer becomes the security conscience of your agent fleet: always watching, interpreting and enabling distributed and trusted autonomy.
As agents continue to proliferate, the ability to deploy them responsibly will define who can scale securely and who introduces unnecessary risk. To secure agent ecosystems, organizations should integrate security from the start of deployment, continuously monitor behavior and access, maintain strong human oversight, and regularly audit and update security policies.
Enterprises that get this right will unlock significant productivity and resilience; not by slowing down agents, but by giving them the security and governance they need to operate safely and responsibly.
We list the best IT Automation software.
This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro
- Amazon has renewed Invincible for a fifth season
- The highly-rated animated show's fourth entry will arrive sometime in 2026
- Matthew Rhys has also joined the Prime Video series' voice cast for season 4
Amazon has revealed Invincible has been renewed for a fifth season.
The announcement, which was made on the adult animated show's social media channels yesterday (July 17), confirms Invincible season 4 won't be its final entry. Considering how popular the critically-acclaimed Prime Video series is, that's no great surprise, but it's nonetheless pleasing to hear that Mark Grayson's story will continue on one of the world's best streaming services.
Season 5, Mark... this is GOOD NEWS pic.twitter.com/Kft2aTuS5SJuly 17, 2025
Interestingly, Amazon also revealed that the show's cast had already completed recording their lines for its fifth season.
It's likely, then, that Invincible season 5 could be ready to go in 2027 and maintain Prime Video's recently established plan to release new seasons annually: Invincible season 2 part 1 aired in late 2023, Invincible season 2 part 2 launched in early 2025, and its third and most recent season released earlier this year. Season 4 is set to arrive in 2026, too, so there's no reason to suspect the show's fifth installment won't arrive a year after that.
Which character might be voiced by Matthew Rhys in Invincible season 4?Welsh actor Matthew Rhys has joined the cast for Invincible's fourth season (Image credit: Michael Loccisano)Invincible's latest renewal and the completion of voice work on season 5 weren't the only announcements made in the above video. Indeed, Amazon also revealed that Matthew Rhys (The Americans, Perry Mason) had joined the voice cast for one of the best Prime Video shows' fourth chapter.
Understandably, Rhys' role is being kept under wraps, but that doesn't mean we can't guess which character he'll portray. After all, there are plenty of candidates in the series' graphic novel namesake who've yet to show up in its animated adaptation.
So, who could Rhys be voicing? Grand Regent Thragg is a possibility but, as much as I like Rhys as an actor, I'm not sure he has the gravitas to play the Viltrum Empire's commander-in-chief. J.K. Simmons, who plays Omni-Man, and Jeffrey Dean Morgan, who voices Conquest, have proven that actors with deeper voices are best suited to portray Viltrumites, so I don't think Rhys is the right fit for Thragg.
If he plays a villain, he might be better hamming it up as someone like Dinosaurus – real name David Anders – who's something of an superhuman eco-terrorist and, as his name implies, a formidable shape-shifting humanoid reptile when he's in his dino-form. Dinosaurus is a dangerous albeit silly character, so Rhys might be the perfect fit to play him.
On the more heroic front, Rhys may be the ideal actor for someone like Space Racer or Tech Jacket. We've briefly seen these superpowered beings in past seasons of Invincible, but neither character has uttered a word yet. It's possible, then, that Rhys has been tapped to voice one of them.
I guess we'll find out for sure when Invincible season 4 is eventually released. In the meantime, read the section below for more coverage on the series' latest season.
You might also like- Invincible season 3 review: another soaring entry of the popular Prime Video series that packs a real punch in more ways than one
- Invincible season 3 ending explained: is [spoiler] dead, Damien Darkblood end credits scene, and more big questions answered
- Invincible season 3 episode 7 makes good on a two-year-old Instagram post and a wild rumor about Jeffrey Dean Morgan
- CISPE members gain access to Microsoft software on a PAYG CSP-Hoster Program
- Privacy and sovereignty benefits also come with the latest update
- It's unclear if this draws a line under the two-year battle
In a major win for European cloud providers, CISPE members can now offer Microsoft software under a pay-as-you-go model via Microsoft CSP-Hoster Program, with licensing costs set to be more closely matched to those offers on Microsoft's own Azure platform.
The news comes after several hurdles in CISPE's battle with the tech giant, but with the group's members now able to host workloads without sharing customer data with Microsoft, the latest development marks a major customer privacy win.
However, some ongoing issues and limitations persist even after two years of battling, including the enforcement of Entra ID with Microsoft 365 – no alternative identity management is allowed.
CISPE wins big against MicrosoftCISPE described the win as a "landmark agreement" that has cost, privacy and sovereignty benefits.
When generally available, Microsoft 365 Local will enable deployment on local (sovereign) European cloud infrastructure, while customers will no longer have to pay more money to access Microsoft software on non-Microsoft platforms.
Despite reaching some agreements over the course of the last two years, CISPE's continued lobbying has produced even more results.
"The agreement we reached with Microsoft marks a significant breakthrough in our long-standing efforts to ensure a level playing field,” CISPE Secretary General Francisco Mingorance explained.
"For enterprise customers, the new programs directly address previous concerns of CISPE members and empowers European enterprises to choose among a wide range of cloud solutions that meet their sovereignty, compliance, and economic needs."
The news comes just days after we learned that Microsoft was readying a counterproposal after failing to meet a previous deadline to settle the saga.
TechRadar Pro has asked Microsoft to comment on the development, but we did not receive an immediate response. We also asked CISPE if this marks the end of the battle.
You might also like- We've listed the best cloud hosting providers
- Microsoft joins European cloud collective following settlement
- Keep your data online with the best cloud storage and best cloud backup tools
- The Pixel Watch 4 is due to make its debut on August 20
- It's apparently going to be more repairable than the Pixel Watch 3
- Two new emergency features are also now rumored
We've got another Made By Google event fast approaching on Wednesday, August 20, where we should get the grand unveiling of the Pixel Watch 4 – and some fresh leaks have now given us a better idea of what's on the way with this wearable.
According to information obtained by Android Headlines, the smartwatch is going to be much more repairable: it "can actually be taken apart and serviced when needed", the report says, unlike the Google Pixel Watch 3, which launched last year.
This isn't a complete surprise, because a Google executive went on the record back in September saying that improved repairability was something the company was "thinking through" when it came to the Pixel Watch series and other Google products.
If this is something that we see happen this year, it would certainly be welcome and could save you from having to pay out for a whole new smartwatch. It's also been an issue that we've had with the Pixel Watch for several years now.
Plus more emergency featuresWe're expecting the Pixel Watch 4 to come in two sizes (Image credit: Future/Lance Ulanoff)Android Headlines is also predicting a couple of emergency features that are apparently heading to the Pixel Watch 4, building on top of existing features such as Emergency SOS, Crash Detection, and Fall Detection.
Added to that list we're going to get Breathing Emergencies, it seems. This will alert you when the Pixel Watch 4 detects that your oxygen levels are low, via the SpO2 sensor – so you can go and get yourself checked out if you need to.
Then there's Emergency Satellite Communications, which we're assuming would be an extension of the feature on the latest Pixel phones. It means you'd be able to contact the emergency services, even if you don't have any cell or Wi-Fi signal.
We don't get too much in the way of detail in this leak, but these are upgrades to look out for at the big show next month. TechRadar will of course be covering all the news live, and the product unveilings should also include the Google Pixel 10 series.
You might also like- A string of code in Samsung beta software mentions the Galaxy Z Trifold's codename
- It also mentions the Snapdragon 8 Elite chipset
- Code or no code, we'd be surprised if this chipset wasn't used
Samsung’s first tri-fold phone is likely to land soon, with recent leaks suggesting that it could launch sometime in October with the name Samsung Galaxy Z Trifold, and now we have a good idea of what chipset might be powering it too.
In a new internal beta of One UI 8 (which is the interface Samsung uses on its Android phones), Android Authority has spotted a string of code that includes ‘siop_q7mq_sm8750.’ That might look meaningless, but reportedly ‘q7mq’ is a codename that has previously been linked to the Galaxy Z Trifold, while ‘sm8750’ is the model number of the Snapdragon 8 Elite chipset.
So putting that together, it certainly looks a lot like the Samsung Galaxy Z Trifold will be powered by the Snapdragon 8 Elite.
The Snapdragon 8 Elite delivers powerful performance (Image credit: Qualcomm)The obvious choiceThis though, isn’t surprising at all – in fact it would be surprising if that chipset wasn’t used. After all, this is the top chipset currently available to Android phones, and it’s also used by the Samsung Galaxy S25 series and the Samsung Galaxy Z Fold 7.
So while it’s not impossible that the Galaxy Z Trifold could use something else, such as perhaps the Samsung-made Exynos 2500 used in the Samsung Galaxy Z Flip 7, it would seem an odd choice for what’s sure to be the company’s priciest phone to use anything other than the best available chipset.
Still, it’s always possible that Samsung could have chosen to keep the cost down by using something else, so it’s still somewhat reassuring to see the Snapdragon 8 Elite mentioned here – even if that does mean this tri-fold phone will probably be exceptionally expensive.
The mention of the Snapdragon 8 Elite also further suggests we’ll see the phone soon, as from late 2025 we’ll probably start seeing the first phones with the Snapdragon 8 Elite 2 – so if Samsung wasn’t launching the Galaxy Z Trifold in the next few months, the company would probably opt for the next generation of this chipset.
But while this code string is compelling evidence, we still can’t be totally sure that the Snapdragon 8 Elite will be used.
We’ll perhaps find out in October – but reports suggest the Samsung Galaxy Z Trifold might initially only be available in China and South Korea, so you may not be able to buy it anyway. If it’s a hit though, we’d expect there will be a more widely available successor.
You might also like- Turtle Beach has revealed a new Donkey Kong-themed Rematch controller
- The controller has a unique lenticular design that shifts when the pad is tilted
- Pre-orders are open now for launch on October 12, 2025
Gaming brand Turtle Beach has unveiled - or should I say, unpeeled - a Donkey Kong-themed version of its popular Rematch controller.
Compatible with the Nintendo Switch and Nintendo Switch 2, this is the latest of many Rematch controllers that sport an officially licensed Nintendo design. This Donkey Kong version differs slightly, though, in that it has a lenticular design with shifting imagery when you tilt the controller. Pretty cool.
Pre-orders are open now for the Donkey Kong-themed Rematch controller on Turtle Beach's website. It costs $59.99 / £49.99, and is currently slated to ship on October 12, 2025. That's a pretty bang average price among the best Nintendo Switch controllers, and something worth considering if you're not up for shelling out $84.99 / £74.99 for a Nintendo Switch 2 Pro Controller.
Otherwise, this Rematch controller isn't much different than others in the brand lineup. Other features include wireless connectivity, gyro aim support, two remappable buttons on the rear, and a claimed battery life of up to 40 hours, similar to the Nintendo Switch 2 Pro Controller.
While compatible with Switch 2, you unfortunately won't be getting a C button here for quick access to the console's GameChat feature. For that, you'll need to access it from the console's Home dashboard instead. Not a big deal, really, but something to keep in mind when you're playing with pals.
You might also like...- Donkey Kong wears clothes in Banaza because Nintendo was 'conscious' of what the ape would look like 'from the back'
- Donkey Kong Bananza doesn't offer one of the Nintendo Switch 2's promised features
- Sega confirms the launch of the physical Nintendo Switch 2 version of Sonic Racing: CrossWorlds in 'early 2026' – and it seemingly won't use the controversial game-key card system
- The Epic Games Store summer sale is here
- Users can claim a free copy of Sid Meier’s Civilization 6: Platinum Edition
- It's available until July 24, 2025
The Epic Games Store summer sale has landed, with savings on some of this year's top titles like Clair Obscur: Expedition 33 and Assassin's Creed Shadows. Shoppers can also claim 20% Epic Rewards on purchases made using Epic's payment system, giving you some left over to spend on even more games.
To celebrate, Epic Games Store users can claim a free PC copy of Sid Meier’s Civilization 6: Platinum Edition which usually costs $79.99 / £68.99.
Considered one of the greatest strategy games of all time, Civilization 6 sees you competing to grow a historical civilization in a board-game like world.
Random map generation and countless playable leaders with their own unique perks and abilities make it a highly replayable game, and one that I have spent more than 260 hours in on PC alone.
This Platinum Edition version is also overflowing with included downloadable content (DLC). You get Vikings Scenario Pack, Poland Civilization & Scenario Pack, Australia Civilization & Scenario Pack, Persia and Macedon Civilization & Scenario Pack, Nubia Civilization & Scenario Pack, and Khmer and Indonesia Civilization & Scenario Pack.
These all add new playable leaders or in-game challenges and, if that wasn't enough, you even get the full Rise and Fall and Gathering Storm expansions. These not only introduce new leaders and challenges, but a wealth of new mechanics that make the experience even more enjoyable and realistic.
The game is only available for free for a limited time, so make sure you snap it up before July 24, 2025.
You might also like...