Only days after OpenAI launched its new Atlas browser, hackers have discovered that it lacks protection from this type of attack.

An ethical hacker demonstrated that ChatGPT Atlas is vulnerable to clipboard injection attacks. Atlas' agent mode might click ...