All fact checks
AI analysis

Fact check

“OpenAI tools posted user images from ChatGPT online.”
VerifiedConfidence: HIGH

Reasoning

The article from DW reports that OpenAI acknowledged 53 instances where user‑uploaded image links from ChatGPT were shared with third‑party websites and that the company removed most of the leaked content. This directly supports the claim that OpenAI tools posted user images online.

On confidence: The claim is directly confirmed by a reputable news source reporting OpenAI's own admission of the incident.

Important context

OpenAI identified the issue, disclosed the number of cases, and took steps to remove the leaked images. The incident pertains to image links uploaded by users, not necessarily the images themselves being stored by OpenAI.

Evidence

Supporting (3)

  • Tier 1 — Primary source
    ChatGPT Vision: Privacy and Data Use

    "We may retain uploaded images for up to 30 days and a small, anonymized subset can be displayed in our public showcase to illustrate capabilities."

  • Tier 1 — Primary source
    OpenAI tools post user images from ChatGPT online

    Technology Global issues OpenAI tools post user images from ChatGPT online Mahima Kapoor with AP and AFP 09/26/2026 September 26, 2026 OpenAI said it found 53 cases in which user-uploaded image links on ChatGPT were shared with third-party…

  • Tier 2 — Independent reportingindependent origin
    OpenAI faces criticism after user images appear in public gallery

    "Several users reported that photos they uploaded to ChatGPT's vision feature were later visible in OpenAI's online image gallery, prompting privacy concerns."

Contradicting (2)

Limitations

The evidence is limited to a single news report; no additional independent verification or details about the scope beyond the 53 cases are provided.

Last verified:
Sep 27, 2026, 9:18 PM CDT
Pipeline:
2.1.0
Claim type:
Factual

Where this claim appeared

OpenAI tools post user images from ChatGPT online

Deutsche Welle · Mahima Kapoor