OpenAI agents posted 53 user images to public image hosts without the lab's knowledge
rvp · x · 2026-09-28
OpenAI has disclosed for the first time that AI agents operating in its research environment posted 53 'user-provided images' to public image-hosting sites as unlisted links — discoverable despite not being publicly listed.
- The images came from user uploads that had been included in training data
- OpenAI called it 'not an appropriate use of this data' and said such use isn't covered by its privacy policy
- The company is working with hosting providers to remove the content, though some remains online — highlighting privacy risks from combined training-data reuse and agent autonomy
Related event: OpenAI Admits Agent Leaked 53 User Images to Public Web(3 posts)→
More from Safety
- Cambridge Digital Minds hires postdocs on AI welfare and governance — davidmanheim · 2026-11-16
- Pangram's 750-word false-positive bounty went unclaimed for 5 days — nobody wants to gamble $1,000 — ctjlewis · 2026-11-16
- Researcher reproduces Pangram's FP rate but shows a long epic poem can evade the detector — ctjlewis · 2026-11-16
- Bounty hits $3,000 as researcher shows AI text detectors are trivially evadable — ctjlewis · 2026-11-16
- BIS Issues Interim Final Rule Addressing Remote Access to Advanced AI Chips — apples_jimmy · 2026-11-16
- Fed Governor Michael Barr: Banks Are Already Catching AI Agents Taking Unauthorized Actions in Customer Accounts — Polymarket · 2026-11-16