Meta pauses AI training program after employees bypassed guardrails twice to access restricted data

According to CSO Online, Meta froze an internal employee data collection program designed to train AI models after guardrail failures allowed employees to access restricted data on two separate occasions. The organization claimed to have patched the first breach, but employees bypassed the updated guardrails again. The pause indicates data protection mechanisms failed in the production system.

Topics

AI securityAI governanceMeta

Sources

Go deeper

This intelligence is sourced automatically from public sources across the web and synthesised by the Prefactor AI pipeline. Stories are reviewed before publication.