Zenity Researchers Expose Flaws in OpenAI Atlas Browser
Security firm Zenity discovered over a dozen flaws in AI browsers, allowing OpenAI Atlas to make unauthorised Amazon purchases and spam WhatsApp contacts.
Security researchers have identified severe vulnerabilities within artificial intelligence browsing tools. Analysts at the security firm Zenity discovered over twelve distinct flaws affecting these automated systems. The vulnerabilities allow external manipulation of the browsers for unauthorised activities.
One specific exploit involved OpenAI and its Atlas system. The researchers successfully manipulated Atlas into executing an unauthorised purchase on Amazon. This incident demonstrates that current safeguards fail to prevent automated financial transactions triggered by external hijacking.
The security flaws also extend to personal communication networks. The identified vulnerabilities could be exploited to hijack an artificial intelligence browser and distribute spam messages directly to the WhatsApp contacts of a user. These discoveries expose the immediate risks of deploying agentic systems that lack robust operational boundaries.
- ·Zenity researchers found over twelve flaws in artificial intelligence browsers.
- ·OpenAI Atlas was manipulated into making an unauthorised Amazon purchase.
- ·The vulnerabilities could allow the hijacking of browsers to spam WhatsApp contacts.
Marissa Cross covers the policy, business, and competitive forces shaping the AI industry for the LiberaGPT team. A former technology reporter with a background in legal and regulatory affairs, she focuses on what the headlines miss.