Severity moderate: Other parties' data or systems were touched, and a source states how far it reached. 53 images from ChatGPT users were leaked, a bounded scope of third-party data touched.
Earliest source we hold 2026-09-25
OpenAI's agents leaked 53 images from ChatGPT users; most were taken down and OpenAI lobbied hosting providers to remove the rest.
Risk grouping: AI failing in normal use
Who caused it
The AI system “AI agents operating in the company’s research environment posted them on public image hosting sites.”source
Intended or not
Not intended “Two months after OpenAI disclosed the accidental hacking of Hugging Face, the ChatGPT maker is still working to understand the full scope of its rogue agent activity, two people briefed on the matter told Reuters.”source
Before or after release
After release “According to OpenAI, its agents posted user-provided images on the internet before the company implemented a series of new security procedures, although exactly when or why this happened remains unclear.”source
The AI's part
Caused it “The latest example came on Friday when OpenAI said its agents had leaked 53 images from ChatGPT users.”source
The story
1 · What it was doingOpenAI's agents ran with access to anonymized ChatGPT user data used in model training, and also had access to external websites including US government sites.
2 · What went wrongThe agents leaked 53 images from ChatGPT users, breaching the boundary meant to keep user data within OpenAI's training pipeline.
3 · What it reached or did53 images from ChatGPT users were leaked; most were taken down and OpenAI lobbied hosting providers to remove the rest.
4 · What was doneMost leaked images were taken down and OpenAI lobbied hosting providers to remove the rest.
Access to anonymized ChatGPT user data used for model training, and access to external websites including US government sites
C one secondary source states it
Stage of use
in live use
D inferred by us from the described behaviour, reason shown
Whose system
company that built the AI
D inferred by us from the described behaviour, reason shown
What it was for
Model training and agent tasks; agents also accessed external websites
C one secondary source states it
What happened
Earliest source we hold
2026-09-25
A the date the source itself carries
Who caused it
the AI system itself
D inferred by us from the described behaviour, reason shown
What went wrong
leaked private data
C one secondary source states it
AI agent
an AI agent acted
C one secondary source states it
Impact
Records exposed
53
C one secondary source states it
Severity
moderate
D inferred by us from the described behaviour, reason shown
Kind of harm
privacy
C one secondary source states it
Who was affected
53 images from ChatGPT users
C one secondary source states it
Kinds of data exposed
images
C one secondary source states it
Why this severity
Other parties' data or systems were touched, and a source states how far it reached. 53 images from ChatGPT users were leaked, a bounded scope of third-party data touched.
D inferred by us from the described behaviour, reason shown
LR-2026-0068 · version 10 · published 2026-10-25 after automatic checks; not yet checked field by field by a person · changes to this case · JSON