News

• Android 16 is causing problems for some Pixel devices
• Navigation buttons and gestures are freezing
• As yet Google hasn't said anything officially about the bugs

Android 16 was officially pushed out to Pixel owners last week, but reports are coming in of the software update causing some pretty major bugs on certain handsets, related to gestures and interface and navigation.

As noted by Android Headlines, the reports say certain navigation buttons and gestures don't respond as they should – meaning certain operations (like going back to the previous screen) take seconds to complete, or don't happen at all.

It's not difficult to find users having problems: on Reddit, we've got posts about unresponsive buttons on a Google Pixel 8, swiping up not doing anything on a Pixel 9 Pro XL, and multiple gestures not functioning on a Pixel 6.

Over on X (formerly Twitter), Artem Russakovskii from Android Police says the back gesture only works intermittently since updating to Android 16 – and there's a screen recording included so you can see what happens (or rather, doesn't happen).

Here's a demo of the back button randomly not working on Android 16.I grabbed a bug report and submitted it to Google engineering along with this reproduction screencast. Hopefully, they'll figure it out. pic.twitter.com/nEmifqQRvbJune 14, 2025

While there are numerous reports of issues online, these problems clearly aren't affecting everyone: since upgrading to Android 16 on my Google Pixel 9 phone, I haven't noticed any problems using the navigation gestures.

As yet, Google hasn't yet said anything officially about these problems. If you are affected, you can report the bug to Google. With so many users now complaining, no doubt this is going to get addressed sooner rather than later.

What makes these bugs particularly annoying is that they relate to a fundamental part of the user experience, and affect multiple apps and screens. Being able to get around your phone is a pretty fundamental part of using it.

Android 16 brings with it a number of useful upgrades, including a new Pixel VIPs feature for your favorite contacts, but we're still waiting for the big Material 3 Expressive revamp to arrive – which should be along later this year.

• Android 16's new battery health tools will skip older Pixel phones
• These are the best Pixel phones you can buy
• Here's how Android 16 is going to be able to install your apps faster

• Fake wallet apps ask for your 12-word phrase and quietly drain your crypto funds
• CRIL found over 20 Play Store apps built solely to steal users’ crypto credentials
• Malicious apps used WebView to fake real login pages from PancakeSwap and others

New research by Cyble Research and Intelligence Labs (CRIL) has uncovered a large-scale phishing campaign involving more than 20 Android applications listed on the Google Play Store.

These apps, which appeared to be legitimate cryptocurrency wallet tools, were created with a singular purpose: stealing users’ mnemonic phrases, the crucial 12-word keys that provide full access to crypto wallets.

Once compromised, victims risk losing their entire cryptocurrency holdings, with no possibility of recovery.

Many of the malicious apps were built using the Median framework, which enables the rapid conversion of websites into Android applications.

Using this method, threat actors embedded phishing URLs directly into the app code or within privacy policy documents.

These links would then load deceptive login pages via a WebView, tricking users into entering their mnemonic phrases under the false belief they were interacting with trusted wallet services such as PancakeSwap, SushiSwap, Raydium, and Hyperliquid.

For example, a fraudulent PancakeSwap app used the URL hxxps://pancakefentfloyd[.]cz/api.php, which led to a phishing page mimicking the legitimate PancakeSwap interface.

Likewise, a fake Raydium app redirected users to hxxps://piwalletblog[.]blog to carry out a similar scam.

Despite variations in branding, these apps shared a common objective: extracting users’ private access keys.

CRIL's analysis revealed that the phishing infrastructure supporting these apps was extensive. The IP address 94.156.177[.]209, used to host these malicious pages, was linked to over 50 other phishing domains.

These domains imitate popular crypto platforms and are reused across multiple apps, indicating a centralized and well-resourced operation.

Some malicious apps were even published under developer accounts previously associated with legitimate software, such as gaming or streaming applications, further lowering user suspicion.

This tactic complicates detection, as even advanced mobile security tools may struggle to identify threats hidden behind familiar branding or developer profiles.

To protect against such attacks, CRIL advises users to download apps only from verified developers and avoid any that request sensitive information.

Using reputable Android antivirus or endpoint protection software, along with ensuring that Google Play Protect is enabled, adds an important, though not infallible, layer of defense.

Strong, unique passwords and multi-factor authentication should be standard practice, and biometric security features should be enabled when available.

Users should also avoid clicking on suspicious links received via SMS or email, and never enter sensitive information into mobile apps unless their legitimacy is certain.

Ultimately, no legitimate app should ever request a full mnemonic phrase through a login prompt. If that happens, it’s likely already too late.

• 1. Pancake Swap
  Package: co.median.android.pkmxaj
  Privacy Policy: hxxps://pancakefentfloyd.cz/privatepolicy.html
• 2. Suiet Wallet
  Package: co.median.android.ljqjry
  Privacy Policy: hxxps://suietsiz.cz/privatepolicy.html
• 3. Hyperliquid
  Package: co.median.android.jroylx
  Privacy Policy: hxxps://hyperliqw.sbs/privatepolicy.html
• 4. Raydium
  Package: co.median.android.yakmje
  Privacy Policy: hxxps://raydifloyd.cz/privatepolicy.html
• 5. Hyperliquid
  Package: co.median.android.aaxblp
  Privacy Policy: hxxps://hyperliqw.sbs/privatepolicy.html
• 6. BullX Crypto
  Package: co.median.android.ozjwka
  Privacy Policy: hxxps://bullxni.sbs/privatepolicy.html
• 7. OpenOcean Exchange
  Package: co.median.android.ozjjkx
  Privacy Policy: hxxps://openoceansi.sbs/privatepolicy.html
• 8. Suiet Wallet
  Package: co.median.android.mpeaaw
  Privacy Policy: hxxps://suietsiz.cz/privatepolicy.html
• 9. Meteora Exchange
  Package: co.median.android.kbxqaj
  Privacy Policy: hxxps://meteorafloydoverdose.sbs/privatepolicy.html
• 10. Raydium
  Package: co.median.android.epwzyq
  Privacy Policy: hxxps://raydifloyd.cz/privatepolicy.html
• 11. SushiSwap
  Package: co.median.android.pkezyz
  Privacy Policy: hxxps://sushijames.sbs/privatepolicy.html
• 12. Raydium
  Package: co.median.android.pkzylr
  Privacy Policy: hxxps://raydifloyd.cz/privatepolicy.html
• 13. SushiSwap
  Package: co.median.android.brlljb
  Privacy Policy: hxxps://sushijames.sbs/privatepolicy.html
• 14. Hyperliquid
  Package: co.median.android.djerqq
  Privacy Policy: hxxps://hyperliqw.sbs/privatepolicy.html
• 15. Suiet Wallet
  Package: co.median.android.epeall
  Privacy Policy: hxxps://suietwz.sbs/privatepolicy.html
• 16. BullX Crypto
  Package: co.median.android.braqdy
  Privacy Policy: hxxps://bullxni.sbs/privatepolicy.html
• 17. Harvest Finance blog
  Package: co.median.android.ljmeob
  Privacy Policy: hxxps://harvestfin.sbs/privatepolicy.html
• 18. Pancake Swap
  Package: co.median.android.djrdyk
  Privacy Policy: hxxps://pancakefentfloyd.cz/privatepolicy.html
• 19. Hyperliquid
  Package: co.median.android.epbdbn
  Privacy Policy: hxxps://hyperliqw.sbs/privatepolicy.html
• 20. Suiet Wallet
  Package: co.median.android.noxmdz
  Privacy Policy: hxxps://suietwz.sbs/privatepolicy.html
• 21. Raydium
  Package: cryptoknowledge.rays
  Privacy Policy: hxxps://www.termsfeed.com/live/a4ec5c75-145c-47b3-8b10-d43164f83bfc
• 22. PancakeSwap
  Package: com.cryptoknowledge.quizzz
  Privacy Policy: hxxps://www.termsfeed.com/live/a4ec5c75-145c-47b3-8b10-d43164f83bfc

• These are the best external hard drives in the market
• Check out the best mini PCs for all budgets
• Build your own superfast virtual 32TB Gen5 SSD with this RAID PCIe Gen5 card

• AI-generated documents are replacing traditional forgeries and bypassing verification faster than ever imagined
• Synthetic identity fraud has surged 195% globally, with Europe and North America hardest hit
• Most fraud now happens after onboarding, using tricks like credential stuffing and device spoofing

Synthetic identity document fraud is now spreading at an alarming pace globally, driven almost entirely by the misuse of generative AI, experts have warned.

New data from Sumsub has claimed the creation of fake documents using AI grew by 195% worldwide between Q1 2024 and Q1 2025.

The most dramatic spike was recorded in Europe, where synthetic ID fraud surged by 378%, followed by North America (311%) and the MENA region (258%).

This trend appears to be gaining ground on traditional document forgery, which has declined sharply in most regions, dropping by 82% in Africa and 50% in North America - with Europe the exception, where traditional forgeries increasing by 33%.

This shift suggests scammers are moving from conventional techniques to AI-generated identities, which can now be produced in seconds using largely unregulated tools.

The industries most affected by this trend include crypto, trading, transportation, fintech, and gambling, all sectors which often rely on rapid onboarding processes and may not yet have strong enough safeguards to detect synthetic documents before damage is done.

Compounding this vulnerability is the fact over three quarter (76%) of fraud now occurs after the onboarding stage, using methods such as device fingerprint spoofing, credential stuffing, and so-called “money muling.”

“The rise of AI-generated fraud is exposing critical flaws in traditional verification systems, and this data shows sophisticated fake documents are becoming a key tool in the fraudster’s arsenal,” said Pavel Goldman-Kalaydin, Head of AI/ML at Sumsub.

“We expect companies like OpenAI to strengthen safeguards controlling the use of its programs to generate passports, IDs, and other personal documents.”

Parallel to this trend is a dramatic rise in deepfake-related fraud, where North America saw a staggering 1,100% increase in such cases, with Europe following at 900%.

Nations like Hong Kong, Singapore, and Germany saw the highest individual surges, raising concerns about the future role of manipulated visual media in identity fraud schemes.

For individuals and organizations trying to protect themselves, relying solely on conventional KYC protocols may no longer be enough. Businesses are advised to use identity management software and multi-layered verification systems for stronger protection.

Consumers should also consider top-rated ID theft protection solutions, especially those that monitor for synthetic identity use. Using reputable people search tools can add an extra layer of safety when verifying unknown contacts or confirming someone’s legitimacy online.

• These are the best VPN with antivirus solutions
• Take a look at our pick of the best password managers
• The AI arms race: why we need AI to fight AI attacks

We’ve known that Cyberpunk 2077 is eventually coming to the Mac, and developers CD Projekt Red (CDPR) have been hard at work – and still are – at getting the title fit to run on any Apple Silicon-powered device. That means the massive AAA title, which has been frequently used to benchmark systems, will be able to run on the M1 chip or newer.

It was reconfirmed at WWDC 2025 – Apple’s week-long developer conference – that it would be arriving later this year, and back with the launch of the new Mac Studio, TechRadar saw a demo of it running on that super-powered computer.

Now, though, I had the chance to get another demo of Cyberpunk 2077: Ultimate Edition – a newer version of the in-development title – on a 16-inch MacBook Pro with the M4 Max chip and 128GB of RAM.

It was on ‘Ultra’ settings with a buttery smooth and excellently rendered 120 frames per second. And yes, this performance is derived from Apple’s M4 Max chip – no discrete GPU required. Further, this was fully taking advantage of Apple's Metal 4, coming later this year, including MetalFX Frame Interpolation.

It looked fantastic and was responsive, but that had me thinking about the plans to launch this on any Apple Silicon Mac.

I was curious about performance on less powerful chips within Apple’s lineup, as the requirement for the title is expected to be Apple Silicon, meaning it should run base M1, M2, M3, or M4 chips as well as the more powerful creations.

I asked Pawel Sasko, Associate Game Director of CD Projekt Red, about what performance might look like on other Macs with different levels of performance from various chips, like even a MacBook Air.

He first explained that Cyberpunk 2077 is running on CDPR’s proprietary Red Engine Four, which is designed to be scalable. “It's trying to use everything that given platform provides,” Sasko explained and called out that the title is available on a variety of devices already, including the Nintendo Switch 2 ahead of its release on the Mac.

"Red engine is basically trying to check what hardware you have and squeeze out the max you can have,” he continued. Meaning that Cyberpunk 2077 will work to run and deliver the best performance based on the chip you have inside. Sasko couldn’t give a specific number yet based on the chip or a potential Mac SKU, but that’s because the game is still not finalized – CDPR is still working on it and polishing it.

Sasko even shared that they’ll work on optimizing performance until the very end, and recently CDPR figured out how to optimize cloth, also known as what a character is wearing, to make it look better.

Suffice to say, he shared that “We're trying to squeeze everything that given hardware gives us,” meaning it will run the best it possibly can on the Mac – which likely means that a more powerful chip, like an M4, M4 Pro, or M4 Max, will deliver the best results.

And the good news is that we’re likely only a few months away from one of the biggest AAA titles arriving on the Mac – the CDPR team wouldn’t give an exact time frame, but Cyberpunk 2077: Ultimate Edition will arrive in 2025.

I also got to see this demo of the title shortly after Apple unveiled macOS 26 Tahoe, which will introduce the Games app. Although CDPR hasn’t announced if it will be integrating with the new hub, it will work with the new ‘Game Overlay.’

This is a sidebar that can be brought to the top right-hand side of the screen, making it easy to enable game mode, adjust brightness or volume, and even switch controllers, as well as connect with friends. You can also opt to push your MacBook further with different energy modes.

• Microsoft makes fun of macOS Tahoe’s Liquid Glass redesign for ripping off Windows Vista – but Apple could have the last laugh
• macOS Tahoe 26 finally adds a much-needed feature that means I may never use Windows 11 again
• Live from WWDC 2025 – TechRadar podcast unpacks that massive iPadOS update and looks through Liquid Glass

One of the biggest differences between being rich and poor is having people to do things for you. When a person moves from being poor to the middle class, they might get a cleaner or take their clothes to a dry cleaner; when they become rich, they might get a driver, a private chef – an entire entourage devoted to making their lives easier.

There’s growing excitement about the rise of AI-powered agents that act on behalf of consumers - not just during product discovery, but right through to purchase. Search-like queries on ChatGPT might only represent 1% of those on Google, but 1% of the colossal global search market is a huge number of searches, with potentially millions in revenue.

Originally conceived as digital concierges to simplify search, these agents are now making actual purchases, and they’re doing it without ever handing control back to the shopper.

Large payments players are laying the groundwork for AI-based commerce. Visa recently launched a Digital Credential Innovation Hub to explore new identity models for agent-based transactions, while Stripe confirmed it is developing secure transaction capabilities for AI agents. And just yesterday, Google revealed plans for an AI agentic checkout for shopping — a move that confirms this shift is no longer speculative, but imminent.

But beyond the optimistic headlines lies a more complicated picture. What happens when an AI agent makes a purchase for you? And more importantly, what can go wrong?

Let’s be clear: there’s no single “standard” for how agentic shopping works – the process is still evolving, and different platforms take different approaches. That said, here’s a common flow we’ve observed in early implementations.

When a consumer uses an AI agent to shop, the process is superficially simple, but technically intricate. First, the user saves their payment card details — including full PAN, CVV, expiry, billing, and delivery addresses — with their chosen AI platform.

Shoppers are unlikely to be buying very inexpensive items like a pizza or very expensive items like a new car. They probably won’t be using it for goods with a heavy visual emphasis, where part of the enjoyment is browsing until something hits you – clothing being the best example. They are likely, at first, to use agentic AI to help them decide between relatively expensive products that are difficult for non-experts to understand: let’s use a good pair of Bluetooth headphones as an example.

The agent, which could be powered by ChatGPT, Google, TikTok Shop, or Amazon’s AI initiatives, uses natural language to respond to a shopper’s request. Just like a shop clerk, it will ask questions to refine results: how much do you want to spend? Do you want over-ear or in-ear headphones? Are there any features like noise cancelling or waterproofing that you need? It can then refine results and present purchase options.

Once the shopper decides a payment process begins that will be mostly invisible to the shopper:

• The shopper stays in the AI interface and never visits the merchant’s site.
• A “Buy” command within the agent UI triggers the agent to autofill the checkout form on the merchant’s site.
• The merchant receives the full card details as if a human shopper were typing them.
• The agent submits the order, and confirmation is sent via both the agent and the merchant.

Critically, the merchant is unlikely to know that they’re dealing with an agent rather than a human. This introduces risks, because if anything goes wrong — an incorrect item, a delivery mix-up, or pricing error - the shopper must resolve it directly with the merchant, even though they never interacted with the merchant’s website themselves.

In other words: don’t talk to me – talk to my agent.

There are several emerging pitfalls already evident:

Security vulnerabilities: In January 2025, Chinese AI platform DeepSeek was hacked, exposing users' stored credentials. The centralization of payment data in AI agents makes them lucrative targets.

Susceptibility to scams: Fraudsters may design sites specifically to trick agents into completing fake checkouts.

Ambiguity in liability: If an agent misplaces an order or inputs incorrect details, it’s unclear whether the AI provider or the consumer bears responsibility.

Poor compatibility:

• It may not support alternative payment types like PayPal, digital wallets or bank transfers (which account for roughly 45% of eCommerce volume in the EU).
• It can’t easily handle additional checkout steps (e.g. seat selection, delivery slots).
• Struggles with card declines, especially in international transactions, where decline rates can be anywhere from 5 to 30%.

In markets like the EU or Japan, legal requirements around Strong Customer Authentication (SCA) mean that consumers must approve each card transaction, making AI-led flows problematic or non-compliant.

Beyond the immediate risks and logistics, the rise of agents raises fundamental questions about the structure of digital commerce.

Will this model gain traction with consumers? It could fizzle like voice commerce and Amazon’s Dash buttons, which failed to take off due to trust and usability issues. Or it could explode, much like the rise of marketplaces or in-app mobile buying. The answer depends on how much value consumers place on convenience, and how well AI agents can overcome trust and control issues.

If AI agents become the preferred interface for eCommerce, the web as we know it may fragment. Why visit a merchant site at all, when your agent can do the work? This shift could drive the development of Model Context Protocols (MCPs) - AI-optimized data layers that replace websites altogether. Some merchants may respond by blocking known agent IPs or designing checkout flows that frustrate automated systems to force direct interaction. Industries like marketing would fundamentally change as it becomes more important to engage with AI agents than human beings.

Meanwhile, platforms like ChatGPT will need to find ways to monetize their newfound influence. That might mean charging merchants referral fees, sparking the emergence of a new SEO-for-AI ecosystem. But such monetization introduces new questions about trust: if your agent is taking commission from merchants, how unbiased are its recommendations?

As we stand on the edge of this transformation, one thing is certain: the infrastructure of digital payments is being rewritten. The question now is whether consumers - and merchants - are ready to follow their agents into this new era.

And I, for one, will be watching closely - either directly, or via my agent.

LINK!

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