OpenAI Warns Prompt Injection Threat Persists
By
Marcus Lee
• 1 min read
OpenAI acknowledges prompt injection attacks cannot be fully eliminated and is using automated AI attackers to harden its Atlas agentic browser against threats.