By Efosa Udinmwen

Copyright techradar

Skip to main content

Tech Radar Pro

Tech Radar Gaming

Close main menu

the business technology experts

België (Nederlands)

Deutschland

North America

US (English)

Australasia

New Zealand

View Profile

Search TechRadar

Expert Insights

Website builders

Web hosting

Best web hosting
Best office chairs
Best website builder
Best antivirus
Expert Insights

Don’t miss these

How GenAI complacency is becoming cybersecurity’s silent crisis

Researcher tricks ChatGPT into revealing security keys – by saying “I give up”

ChatGPT Agent shows that there’s a whole new world of AI security threats on the way we need to worry about

Asking ChatGPT to help with your security qualms could be putting your data at serious risk

Thousands of organizations have a new, unexpected ’employee’ onboard – and it could be their single biggest security risk

Masked, not erased: how broken redaction fuels AI data leaks

Agentic AI: the rising threat that demands a human-centric cybersecurity response

Cyber Crime
5 worrying ways AI is being used by cybercriminals to target millions of victims

Tackling Shadow AI: how UK businesses can mitigate the risks

Meta patches worrying security bug which could have exposed user AI prompts and responses – and pays the bug hunter $10,000

SMBs are being hit by malicious productivity tools – Zoom and ChatGPT spoofed by hackers

The AI Triple Threat: mitigating the dangers of AI adoption with identity security

AI LLMs are now so clever that they can independently plan and execute cyberattacks without human intervention — and I fear that it is only going to get worse

ChatGPT and other AI tools could be putting users at risk by getting company web addresses wrong

Criminals and scammers are using hacked websites and expired domain names to ‘poison’ ChatGPT with spammy recommendations – here’s how to stay safe

ChatGPT hit by a zero-click, server-side vulnerability that criminals can use to siphon sensitive data – here’s how to stay safe

Efosa Udinmwen

22 September 2025

Attackers can manipulate AI-driven workflows in unanticipated ways

When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works.

(Image credit: NurPhoto / Getty Images)

ChatGPT server-side flaw allows attackers to steal data without any user interaction
ShadowLeak bypasses traditional endpoint security entirely
Millions of business users could be exposed due to ShadowLeak exploits

Enterprises are increasingly using AI tools such as ChatGPT’s Deep Research agent to analyze emails, CRM data, and internal reports for strategic decision-making, experts have warned.

These platforms offer automation and efficiency but also introduce new security challenges, particularly when sensitive business information is involved.
Radware recently revealed a zero-click flaw in ChatGPT’s Deep Research agent, dubbed “ShadowLeak,” but unlike traditional vulnerabilities, this flaw exfiltrates sensitive data covertly.

You may like

How GenAI complacency is becoming cybersecurity’s silent crisis

Researcher tricks ChatGPT into revealing security keys – by saying “I give up”

ChatGPT Agent shows that there’s a whole new world of AI security threats on the way we need to worry about

ShadowLeak: a zero-click, server-side exploit
It allows attackers to exfiltrate sensitive data entirely from OpenAI servers, without requiring any interaction from users.

“This is the quintessential zero-click attack,” said David Aviv, chief technology officer at Radware.
“There is no user action required, no visible cue, and no way for victims to know their data has been compromised. Everything happens entirely behind the scenes through autonomous agent actions on OpenAI cloud servers.”
ShadowLeak also operates independently of endpoints or networks, making detection extremely difficult for enterprise security teams.

Are you a pro? Subscribe to our newsletter
Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!
Contact me with news and offers from other Future brandsReceive email from us on behalf of our trusted partners or sponsorsBy submitting your information you agree to the Terms & Conditions and Privacy Policy and are aged 16 or over.
The researchers demonstrated that simply sending an email with hidden instructions could trigger the Deep Research agent to leak information autonomously.
Pascal Geenens, director of cyber threat intelligence at Radware, explained that “Enterprises adopting AI cannot rely on built-in safeguards alone to prevent abuse.
“AI-driven workflows can be manipulated in ways not yet anticipated, and these attack vectors often bypass the visibility and detection capabilities of traditional security solutions.”

You may like

How GenAI complacency is becoming cybersecurity’s silent crisis

Researcher tricks ChatGPT into revealing security keys – by saying “I give up”

ChatGPT Agent shows that there’s a whole new world of AI security threats on the way we need to worry about

The vulnerability represents the first purely server-side zero-click data exfiltration, leaving almost no evidence from the perspective of businesses.
With ChatGPT reporting over 5 million paying business users, the potential scale of exposure is substantial.
Human oversight and strict access controls remain critical when sensitive data is connected to autonomous AI agents.
Therefore, organizations adopting AI must approach these tools with caution, continuously evaluate security gaps, and combine technology with informed operational practices.
How to stay safe

Implement layered cybersecurity defenses to protect against multiple types of attacks simultaneously.
Regularly monitor AI-driven workflows to detect unusual activity or potential data leaks.
Deploy the best antivirus solutions across systems to protect against traditional malware attacks.
Maintain robust ransomware protection to safeguard sensitive information from lateral movement threats.
Enforce strict access controls and user permissions for AI tools interacting with sensitive data.
Ensure human oversight when autonomous AI agents access or process sensitive information.
Implement logging and auditing of AI agent activity to identify anomalies early.
Integrate additional AI tools for anomaly detection and automated security alerts.
Educate employees on AI-related threats and the risks of autonomous agent workflows.
Combine software defenses, operational best practices, and continuous vigilance to reduce exposure.
You might also like

These are the best firewall offerings around today
These are the best VPNs with antivirus you can use right now
Microsoft announces “world’s most powerful data center” in billion-dollar AI splurge

Efosa Udinmwen

Freelance Journalist

Efosa has been writing about technology for over 7 years, initially driven by curiosity but now fueled by a strong passion for the field. He holds both a Master’s and a PhD in sciences, which provided him with a solid foundation in analytical thinking. Efosa developed a keen interest in technology policy, specifically exploring the intersection of privacy, security, and politics. His research delves into how technological advancements influence regulatory frameworks and societal norms, particularly concerning data protection and cybersecurity. Upon joining TechRadar Pro, in addition to privacy and technology policy, he is also focused on B2B security products. Efosa can be contacted at this email: udinmwenefosa@gmail.com

You must confirm your public display name before commenting

Please logout and then login again, you will then be prompted to enter your display name.

How GenAI complacency is becoming cybersecurity’s silent crisis

Researcher tricks ChatGPT into revealing security keys – by saying “I give up”

ChatGPT Agent shows that there’s a whole new world of AI security threats on the way we need to worry about

Asking ChatGPT to help with your security qualms could be putting your data at serious risk

Thousands of organizations have a new, unexpected ’employee’ onboard – and it could be their single biggest security risk

Masked, not erased: how broken redaction fuels AI data leaks

Latest in Security

Hackers are using GPT-4 to build a virtual assistant – here’s what we know

Scammers build fake FBI crime reporting portals to steal personal info – warns FBI

Ransomware hackers could be targeting GoAnywhere MFT once again – here’s what we know

EU says ransomware to blame for attack which caused chaos at airports

CISA flags some more serious Ivanti software flaws, so patch now

This serious Microsoft Entra flaw could have let hackers infiltrate any user, so patch now

Latest in News

Windows 11 could bring back an old feature for wallpapers from Windows Vista – and it’s about time

How to watch British Open snooker on ITVX (it’s free)

New Windows 11 25H2 update is about to land on your PC – but where’s the excitement?

First trailer for The Mandalorian & Grogu reveals Sigourney Weaver’s mystery Star Wars character, Rotta the Hutt, and the return of The Rise of Skywalker’s best creature

Montblanc just released an e-notebook, and yes it’s staggeringly expensive

PureVPN Linux apps found to leak IPv6 traffic and mess with your firewall – here’s how to secure your data

LATEST ARTICLES

Hackers can now inject AI deepfakes directly into iOS video calls using this tool – here’s how to stay safe

Huawei is planning a 256-core CPU monster to take on AMD EPYC and Intel Xeon range but it won’t land till 2028 – at least that’s the official line

First trailer for The Mandalorian & Grogu reveals Sigourney Weaver’s mystery Star Wars character, Rotta the Hutt, and the return of The Rise of Skywalker’s best creature

Windows 11 could bring back an old feature for wallpapers from Windows Vista – and it’s about time

I review home gadgets for a living, and this air circulator fan is hands-down the best thing I’ve tested this year – here’s why

TechRadar is part of Future US Inc, an international media group and leading digital publisher. Visit our corporate site.

Contact Future’s experts

Terms and conditions

Privacy policy

Cookies policy

Advertise with us

Web notifications

Accessibility Statement

Future US, Inc. Full 7th Floor, 130 West 42nd Street,

Please login or signup to comment

Please wait…