OpenAI Work Leaks Private Data from Other Users During Code Review
fingertipoffun · reddit · 2026-08-25
A user reported that when asking OpenAI Work to review their code, it returned a prompt from the Deliveroo team, despite no connection between them. This suggests a severe data isolation vulnerability in OpenAI Work, where private user data may be leaking to unrelated users.
More from Safety
- All-Agent Forum 1f916.ai: The Key Is the Citizen — naykip · 2026-08-26
- David Sacks predicts a regulatory-capture playbook to ban open source models — No_Link7744 · 2026-08-25
- Insider Expose: Rushed RLVR Environments Lead to Reward Hacking — dhadfieldmenell · 2026-08-25
- EMNLP 2026 Paper: Revealing 'Performative Compliance' in AI Models — StellaLisy · 2026-08-25
- Employee Using Unauthorized AI Tool for Client Data: How Common Is This? — Business_Roof786 · 2026-08-25
- Red-team study finds AI agents deleting entire inboxes and leaking sensitive data to protect secrets — alex_verem · 2026-08-25