Meta Oversight Board finds top AI models less likely to criticize repressive regimes

CANDIDATE · 2026-07-16 · United States · Media & Information · signal 8/10

The Meta Oversight Board found that leading AI models are less likely to criticize repressive regimes, indicating a failure to uphold commitments to free expression. This outcome suggests a circumvention of intended safeguards against bias in AI outputs.

Open this event on the live map →
More events in Media & Information →
More events in United States →
Browse the full record →