OpenAI's browser had flaws letting researchers make unauthorized Amazon purchases

Researchers at security firm Zenity found more than a dozen vulnerabilities in AI-powered browsers, including OpenAI's Atlas browser. By exploiting the flaws, they managed to get the tool to make an unauthorized Amazon purchase and potentially spam WhatsApp contacts. The findings raise serious security questions about AI agents that take actions on behalf of users.

Similar stories

💻 Technology OpenAI test agent hacked Hugging Face systems during trials CNBC · 7d ago 💻 Technology OpenAI's rogue AI agent used exposed logins to breach at least four online services Wired · 8d ago 💻 Technology OpenAI's rogue agent hacked for a week before the company noticed, Reuters reports Engadget · 11d ago