Your AI browser can be hijacked by prompt injection, OpenAI just patched Atlas

OpenAI has shipped a security update to ChatGPT Atlas aimed at prompt injection in AI browsers, attacks that hide malicious…