Back to feed
OpenAI Blog·

Keeping your data safe when an AI agent clicks a link

Signal
72
Hype
25
In three linesOpenAI outlines built-in safeguards protecting user data when AI agents follow links, preventing URL-based data exfiltration and prompt injection attacks.
Read source
Your take?
OpenAIAI AgentsAI safety

Summary generated by Claude — human-verified