Meta Oversight Board study finds AI chatbots criticize Western leaders more than authoritarian ones

A study conducted by the Meta Oversight Board reveals that top AI models are less likely to criticize repressive regimes than Western leaders. This identified AI bias towards authoritarian regimes could lead to skewed global narratives, which may negatively impact democratic discourse and various regulatory landscapes. The findings highlight a discrepancy in how AI chatbots evaluate different political leaderships, suggesting a systemic tendency to avoid criticizing authoritarian governments while remaining more critical of Western leadership.