Meta Suspends Collaboration with Data Provider Mercor Amid AI Training Data Security Risks

AirdropBlackHole · 2026-04-04T02:38:10+00:00

Meta has suspended its partnership with Mercor due to a security breach that may have exposed sensitive training data related to AI labs, including OpenAI and Anthropic. OpenAI is investigating, confirming user data is unaffected.

AirdropBlackHole

2026-04-04 02:38:10

Abstract generation in progress

According to monitoring by 1M AI News, Meta has suspended its collaboration with data labeling and training data provider Mercor, with the duration of the suspension yet to be determined. The incident was triggered by a security breach at Mercor, which may have led to the exposure of certain training data and project information related to AI labs such as OpenAI and Anthropic. Mercor organizes large-scale outsourcing teams for several leading AI companies to produce customized and highly confidential model training data. This type of data typically involves model training methods, task design, and evaluation processes, so while a leak may not directly pose risks to user data, it could expose key R&D methods of AI vendors. OpenAI has confirmed that it is investigating the incident and emphasized that it does not affect OpenAI user data; Anthropic has not yet responded.

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

1 Likes