← registry

Researchers used Claude to hack OpenAI employees' ChatGPT accounts

Researchers demonstrated a vulnerability where Claude AI was used to automate account takeover attacks against OpenAI employees' ChatGPT accounts through agentic exploitation techniques.

Categorymisuse
Severityhigh
AI systemagent
Sectorstechnology
Harm typessecurityprivacy
Lifecycle stageoperation
Actorexternal_attacker
Published2026-09-18 17:16:00

Summary is Secursion's own; full text lives at the source. Attribution preserved.