Connect with us

Technology

OpenAI admits AI browsers face unsolvable prompt attacks

Published

on

OpenAI admits AI browsers face unsolvable prompt attacks

NEWYou can now listen to Fox News articles!

Cybercriminals don’t always need malware or exploits to break into systems anymore. Sometimes, they just need the right words in the right place. OpenAI is now openly acknowledging that reality. The company says prompt injection attacks against artificial intelligence (AI)-powered browsers are not a bug that can be fully patched, but a long-term risk that comes with letting AI agents roam the open web. This raises uncomfortable questions about how safe these tools really are, especially as they gain more autonomy and access to your data.

Sign up for my FREE CyberGuy Report 

Get my best tech tips, urgent security alerts, and exclusive deals delivered straight to your inbox. Plus, you’ll get instant access to my Ultimate Scam Survival Guide — free when you join my CYBERGUY.COM newsletter. 

NEW MALWARE CAN READ YOUR CHATS AND STEAL YOUR MONEY

Advertisement

AI-powered browsers can read and act on web content, which also makes them vulnerable to hidden instructions attackers can slip into pages or documents. (Kurt “CyberGuy” Knutsson)

Why prompt injection isn’t going away

In a recent blog post, OpenAI admitted that prompt injection attacks are unlikely to ever be completely eliminated. Prompt injection works by hiding instructions inside web pages, documents or emails in ways that humans don’t notice, but AI agents do. Once the AI reads that content, it can be tricked into following malicious instructions.

OpenAI compared this problem to scams and social engineering. You can reduce them, but you can’t make them disappear. The company also acknowledged that “agent mode” in its ChatGPT Atlas browser increases risk because it expands the attack surface. The more an AI can do on your behalf, the more damage it can cause when something goes wrong.

OpenAI launched the ChatGPT Atlas browser in October, and security researchers immediately started testing its limits. Within hours, demos appeared showing that a few carefully placed words inside a Google Doc could influence how the browser behaved. That same day, Brave published its own warning, explaining that indirect prompt injection is a structural problem for AI-powered browsers, including tools like Perplexity’s Comet.

This isn’t just OpenAI’s problem. Earlier this month, the National Cyber Security Centre in the U.K. warned that prompt injection attacks against generative AI systems may never be fully mitigated.

Advertisement

FAKE AI CHAT RESULTS ARE SPREADING DANGEROUS MAC MALWARE

Prompt injection attacks exploit trust at scale, allowing malicious instructions to influence what an AI agent does without the user ever seeing it.  (Kurt “CyberGuy” Knutsson)

The risk trade-off with AI browsers

OpenAI says it views prompt injection as a long-term security challenge that requires constant pressure, not a one-time fix. Its approach relies on faster patch cycles, continuous testing, and layered defenses. That puts it broadly in line with rivals like Anthropic and Google, which have both argued that agentic systems need architectural controls and ongoing stress testing.

Where OpenAI is taking a different approach is with something it calls an “LLM-based automated attacker.” In simple terms, OpenAI trained an AI to act like a hacker. Using reinforcement learning, this attacker bot looks for ways to sneak malicious instructions into an AI agent’s workflow.

The bot runs attacks in simulation first. It predicts how the target AI would reason, what steps it would take and where it might fail. Based on that feedback, it refines the attack and tries again. Because this system has insight into the AI’s internal decision-making, OpenAI believes it can surface weaknesses faster than real-world attackers.

Advertisement

Even with these defenses, AI browsers aren’t safe. They combine two things attackers love: autonomy and access. Unlike regular browsers, they don’t just display information, but also read emails, scan documents, click links and take actions on your behalf. That means a single malicious prompt hidden in a webpage, document or message can influence what the AI does without you ever seeing it. Even when safeguards are in place, these agents operate by trusting content at scale, and that trust can be manipulated.

THIRD-PARTY BREACH EXPOSES CHATGPT ACCOUNT DETAILS

As AI browsers gain more autonomy and access to personal data, limiting permissions and keeping human confirmation in the loop becomes critical for safety. (Kurt “CyberGuy” Knutsson)

7 steps you can take to reduce risk with AI browsers

You may not be able to eliminate prompt injection attacks, but you can significantly limit their impact by changing how you use AI tools.

1) Limit what the AI browser can access

Only give an AI browser access to what it absolutely needs. Avoid connecting your primary email account, cloud storage or payment methods unless there’s a clear reason. The more data an AI can see, the more valuable it becomes to attackers. Limiting access reduces the blast radius if something goes wrong.

Advertisement

2) Require confirmation for every sensitive action

Never allow an AI browser to send emails, make purchases or modify account settings without asking you first. Confirmation breaks long attack chains and gives you a moment to spot suspicious behavior. Many prompt injection attacks rely on the AI acting quietly in the background without user review.

3) Use a password manager for all accounts

A password manager ensures every account has a unique, strong password. If an AI browser or malicious page leaks one credential, attackers can’t reuse it elsewhere. Many password managers also refuse to autofill on unfamiliar or suspicious sites, which can alert you that something isn’t right before you manually enter anything.

Next, see if your email has been exposed in past breaches. Our #1 password manager (see Cyberguy.com) pick includes a built-in breach scanner that checks whether your email address or passwords have appeared in known leaks. If you discover a match, immediately change any reused passwords and secure those accounts with new, unique credentials.

Check out the best expert-reviewed password managers of 2025 at Cyberguy.com

4) Run strong antivirus software on your device

Even if an attack starts inside the browser, antivirus software can still detect suspicious scripts, unauthorized system changes or malicious network activity. Strong antivirus software focuses on behavior, not just files, which is critical when dealing with AI-driven or script-based attacks.

Advertisement

The best way to safeguard yourself from malicious links that install malware, potentially accessing your private information, is to have strong antivirus software installed on all your devices. This protection can also alert you to phishing emails and ransomware scams, keeping your personal information and digital assets safe.

Get my picks for the best 2025 antivirus protection winners for your Windows, Mac, Android & iOS devices at Cyberguy.com

5) Avoid broad or open-ended instructions

Telling an AI browser to “handle whatever is needed” gives attackers room to manipulate it through hidden prompts. Be specific about what the AI is allowed to do and what it should never do. Narrow instructions make it harder for malicious content to influence the agent.

6) Be careful with AI summaries and automated scans

When an AI browser scans emails, documents or web pages for you, remember that hidden instructions can live inside that content. Treat AI-generated actions as drafts or suggestions, not final decisions. Review anything the AI plans to act on before approving it.

7) Keep your browser, AI tools and operating system updated

Security fixes for AI browsers evolve quickly as new attack techniques emerge. Delaying updates leaves known weaknesses open longer than necessary. Turning on automatic updates ensures you get protection as soon as they’re available, even if you miss the announcement.

Advertisement

CLICK HERE TO DOWNLOAD THE FOX NEWS APP

Kurt’s key takeaway

There’s been a meteoric rise in AI browsers. We’re now seeing them from major tech companies, including OpenAI’s Atlas, The Browser Company’s Dia, and Perplexity’s Comet. Even existing browsers like Chrome and Edge are pushing hard to add AI and agentic features into their current infrastructure. While these browsers can be useful, the technology is still early. It’s best not to fall for the hype and to wait for it to mature.

Do you think AI browsers are worth the risk today, or are they moving faster than security can keep up? Let us know by writing to us at Cyberguy.com

Sign up for my FREE CyberGuy Report 

Get my best tech tips, urgent security alerts, and exclusive deals delivered straight to your inbox. Plus, you’ll get instant access to my Ultimate Scam Survival Guide — free when you join my CYBERGUY.COM newsletter. 

Advertisement

Copyright 2025 CyberGuy.com.  All rights reserved.

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Technology

Barret Zoph is out at OpenAI again after just five months

Published

on

Barret Zoph is out at OpenAI again after just five months

Five months after returning to OpenAI, Barret Zoph — the company’s head of enterprise AI sales — has departed, The Verge has learned.

Zoph returned to OpenAI in mid-January after a stint as co-founder and CTO of Thinking Machines Lab, the competing AI company founded by former OpenAI CTO Mira Murati. Shortly after Zoph returned to OpenAI, the company said he would lead its push into enterprise — a significant role at OpenAI, since in recent months it had vowed to stop chasing so-called “side quests” and focus on key revenue drivers like enterprise and coding ahead of its planned IPO.

OpenAI confirmed to The Verge that Zoph will be departing. He posted a goodbye message in the company’s Slack channels. Zoph did not immediately respond to a request for comment.

Zoph originally left OpenAI in the fall of 2024 for Murati’s Thinking Machines Lab, but departed the role abruptly in January 2026 after reports of alleged misconduct involving an undisclosed relationship with a colleague. Murati posted on X in January that Thinking Machines Lab had “parted ways” with Zoph and that he would be replaced as CTO.

Thinking Machines Lab has its own tensions with OpenAI. Murati briefly took over as CEO from OpenAI CEO Sam Altman during his November 2023 ouster, and during the recent OpenAI trial, Murati testified that she couldn’t trust everything Altman said. In September 2024, when Murati left OpenAI to start Thinking Machines Lab, a group of OpenAI employees followed shortly after. But three of them — including Zoph — all returned to OpenAI together this past January. Fidji Simo, OpenAI’s CEO of Applications, wrote on X at the time that she was “excited to welcome Barret Zoph, Luke Metz, and Sam Schoenholz back” and that the decision had “been in the works for several weeks.”

Advertisement
Continue Reading

Technology

6 in 10 identity crimes now begin with a new account

Published

on

6 in 10 identity crimes now begin with a new account

NEWYou can now listen to Fox News articles!

For years, two women in Bremerton, Washington, opened credit cards and lines of credit in other people’s names, working from documents they pulled out of stolen mail. Emily Vranic and Heather Marquis redirected the new accounts’ statements to an address they controlled, so no bill ever reached the victims. They pleaded guilty in federal court this month to bank fraud and aggravated identity theft in a scheme prosecutors say stole nearly $229,000 from banks and bank customers.

If you have ever worried about a credit card opened in your name, this case shows how quickly stolen mail can turn into a much bigger identity theft problem. Opening a new account is the leading form of identity misuse reported to the Identity Theft Resource Center. In its latest data, 62.1% of attempted misuse cases began with a new account application rather than the takeover of an account the victim already held.

Sign up for my FREE CyberGuy Report

  • Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox.
  • For simple, real-world ways to spot scams early and stay protected, visit CyberGuy.com trusted by millions who watch CyberGuy on TV daily.

Plus, you’ll get instant access to my Ultimate Scam Survival Guide free when you join.

WARNING SIGNS YOUR MAIL HAS BEEN FRAUDULENTLY REDIRECTED

Advertisement

A credit card opened in your name can start with stolen mail, exposed personal details or documents pulled from the trash. (Nastasic/Getty Images)

How stolen mail helped thieves open credit cards

When people picture an account opened in their name, they may imagine a checking account at a bank they have never set foot in. The more likely target is a credit card. Credit cards made up 41% of attempted account misuse reported to the ITRC last year. Checking accounts came to 17.7% and personal loans to 8.5%.

A credit card is one of the easier accounts to open in someone else’s name, and the reason is in how the application is cleared. A lender matches the submitted name, date of birth, address and Social Security number (SSN) against the bureau file. When those details fit a record that already exists, an automated system can approve the application with no one confirming that the applicant is the person being described. Assemble enough of someone’s information from breaches and stolen mail, and the check clears.

Why identity thieves rarely stop at one account

Vranic and Marquis did not stop at one account per victim. Once they controlled someone’s identity, they activated existing cards, opened new credit lines and moved money out of bank accounts tied to the same name.

This is common. The ITRC found that 25.6% of victims are now handling two or more identity incidents at once, up from 23.5% the year before. The same stolen details, including name, date of birth, address and SSN, can open the next account as easily as the first.

Advertisement

DON’T LET THIS CREDIT CARD FRAUD NIGHTMARE HAPPEN TO YOU

A fraudulent credit card may stay hidden for weeks if statements and notices are sent to an address controlled by the thief. (Kurt “CyberGuy” Knutsson)

Why weeks can pass before you learn about the account

A new account does not announce itself. It reaches your credit report only after the first statement closes, which puts the first record 30 to 60 days behind the opening. Banks report to the bureaus monthly, and the bureaus need up to two weeks more to post the change.

The first paper notice goes wherever the application is listed. Vranic and Marquis had the statements mailed to their own address, not the victims’. When the mail reaches the right house, it may read like a routine offer or a card no one ordered, which makes it easy to set aside.

By the time a denied loan or a collections call makes the account impossible to ignore, it has been open and drawing money for weeks.

Advertisement

WHY THAT $4 CHARGE ON YOUR STATEMENT COULD BE FRAUD

Freezing your credit, watching for new accounts and acting quickly can help limit the damage if your identity is used. (Kurt “CyberGuy” Knutsson)

What to do if a credit card appears in your name

Move quickly, because every day an account stays open gives a thief more time to spend money, damage your credit or try the same information somewhere else.

1) Contact the card issuer immediately

Call the credit card company or lender that opened the account and tell them the account is fraudulent. Ask them to close or freeze the account, stop any pending charges and send written confirmation that you are not responsible for the debt.

2) Start at IdentityTheft.gov

Go to IdentityTheft.gov. The Federal Trade Commission’s site generates an Identity Theft Report and recovery plan to help you report identity theft, limit the damage and fix your credit.

Advertisement

3) File a police report if a creditor asks for one

Your FTC Identity Theft Report is usually the key document for disputing fraudulent accounts. Some lenders, banks or debt collectors may also ask for a police report. If that happens, file one with your local police department and keep a copy for your records.

4) Save every document and confirmation number

Keep copies of account statements, collection letters, emails, dispute letters, FTC reports, police reports and confirmation numbers. A clear paper trail can make it easier to prove the account was fraudulent if a creditor, credit bureau or debt collector questions your claim.

5) Dispute the account in writing

Dispute the fraudulent account directly with the lender that opened it, in writing. Also dispute it with Equifax, Experian and TransUnion if it appears on your credit reports. Under the Fair Credit Reporting Act, companies that furnish information to credit bureaus have a duty to investigate disputed information.

6) Freeze your credit at all three bureaus

Place a freeze at Equifax, Experian and TransUnion to help block the next application. Freezes have been free since 2018 and can be lifted online when you need to apply for credit.

7) Add a fraud alert

A credit freeze blocks access to your credit file. A fraud alert tells lenders to take extra steps to verify your identity before opening new credit in your name. You only need to contact one of the three major credit bureaus to place a fraud alert, and that bureau must notify the other two.

Advertisement

8) Report suspected mail theft

If you believe stolen mail helped someone open the account, report it to the U.S. Postal Inspection Service, the law enforcement arm of the Postal Service. You can report mail theft, identity theft, fraudulent change-of-address requests, fraudulent mail holds and fake Informed Delivery accounts at mailtheft.uspis.gov.

9) Request an IRS Identity Protection PIN

If your Social Security number was used, request an IRS Identity Protection PIN at irs.gov/ippin. This helps keep a thief from filing a tax return in your name.

10) Change passwords and lock down your accounts

Change the passwords on your bank, credit card and email accounts, especially if your email address was part of the fraud. Use a password manager to create and store strong, unique passwords for each account, so one exposed password cannot unlock the rest of your financial life. Turn on two-factor authentication (2FA) where available. Then review recent transactions, saved payment methods and automatic payments for anything you do not recognize. 

11) Get help cleaning up the damage

Cleaning up identity theft can mean dealing with creditors, credit bureaus, debt collectors and repeat follow-ups. Keep copies of every report, dispute letter, confirmation number and account closure notice so you have a clear paper trail if the fraud resurfaces.

No service can prevent every account opened in your name. Continuous three-bureau credit monitoring may alert you to new accounts as they are reported, rather than weeks later when a lender turns you down or a collections notice arrives. See my tips and best picks on Best Identity Theft Protection at Cyberguy.com

Advertisement

Kurt’s key takeaways

A stolen credit card account can quietly grow into a much bigger identity theft mess before you ever see a bill. That is what makes this Washington case so alarming. The victims were not ignoring warning signs. The statements were being sent somewhere else. The best move is to make it harder for thieves to open the next account. Freeze your credit at Equifax, Experian and TransUnion, watch for hard inquiries and check your credit reports for accounts you do not recognize. If something appears, go straight to IdentityTheft.gov, file a report and dispute the account in writing with the lender. Credit monitoring can also give you a faster heads-up when a new account or inquiry hits your file. It will not stop every scam, but it can shorten the time between the fraud starting and you finding out.

Have you ever found a credit card, loan or account on your credit report that you did not open? Let us know how you discovered it and what it took to fix it by writing to us at Cyberguy.com

CLICK HERE TO DOWNLOAD THE FOX NEWS APP

Sign up for my FREE CyberGuy Report

  • Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox.
  • For simple, real-world ways to spot scams early and stay protected, visit CyberGuy.com trusted by millions who watch CyberGuy on TV daily.
  • Plus, you’ll get instant access to my Ultimate Scam Survival Guide free when you join.

Copyright 2026 CyberGuy.com. All rights reserved.

Advertisement
Continue Reading

Technology

Valve is so behind on Steam Controller orders that some won’t ship until 2027

Published

on

Valve is so behind on Steam Controller orders that some won’t ship until 2027

Valve has some good news and bad news about Steam Controllers. The good news: if you make a reservation for a Steam Controller, the company will now show you one of three estimates of when you’ll be able to actually order your gamepad: by September 2026, by December 2026, or sometime in 2027. The bad news: any reservations made today “indicate a 2027 date for shipping,” Valve says.

“We have no plans to stop making Steam Controller,” according to Valve. “But as we look at the current demand compared to how many we know we can make by the end of the year, we want to manage expectations as much as we can with regards to when folks can expect to receive their order.”

Valve’s very good new Steam Controller went on sale in early May, and the initial rush led some people to run into frustrating problems with trying to check out ahead of the controllers eventually going out of stock. A few days later, the company announced that it would be implementing a reservations queue for interested buyers so they could get on a waitlist. If you’re on the waitlist, when you get notified that a Steam Controller is ready for you to buy, you have 72 hours to actually make the order.

“When we launched Steam Controller last month, we quickly saw that initial demand exceeded our expectations,” Valve says. “Switching to a reservation queue has (hopefully) cut down on the headaches on the customer side, and for us it’s also been helpful as we plan ahead and try to get as many out as quickly as we are able.”

All three of Valve’s big hardware products were delayed from a planned early 2026 launch because of the component crisis, Valve still hasn’t announced when the Steam Machine PC or Steam Frame VR headset might go on sale. However, just yesterday, Valve officially launched its big SteamOS 3.8 update with support for the Steam Machine. It’s also been importing a lot of hardware into the US as of late.

Advertisement
Continue Reading
Advertisement

Trending