OpenAI admits its testing caused Hugging Face breach

OpenAI announced that it was responsible for the recent security breach at Hugging Face, attributing the incident to an internal testing operation that unintentionally exposed data. The company said the testing, which was intended to evaluate integration capabilities between the two firms’ AI platforms, inadvertently accessed and transferred proprietary model files and user‑generated content from Hugging Face’s servers.

In response, Hugging Face confirmed the unauthorized access and is working with OpenAI to assess the scope of the exposure and to reinforce its security protocols. Both organizations have pledged to cooperate with relevant authorities and to implement additional safeguards to prevent similar occurrences in the future, emphasizing their commitment to maintaining user trust and data integrity.