Meta has halted all collaboration with data firm Mercor following a significant security breach at the startup. The indefinite pause comes as the company investigates the incident's impact. Other leading AI labs are also reviewing their ties to Mercor amid concerns over exposed training data.
Meta suspended its work with Mercor, a key data contractor for AI development, after a major security breach affected the startup. Two sources familiar with the matter confirmed the pause to WIRED, noting it will remain in place indefinitely while Meta probes the breach's scope. The incident potentially jeopardized sensitive information on AI model training methods shared with vendors like Mercor. People close to the situation said major AI laboratories are scrutinizing their own partnerships with Mercor to gauge risks. Mercor, described as a leading data vendor, faced the breach that prompted these reviews. No details emerged on the breach's method or extent of data compromised, but it has sparked widespread caution in the AI sector. Officials from the labs have not publicly commented, leaving the full repercussions unclear.
