AI as the Perfect Propaganda Machine

The Meta Oversight Board has warned that modern AI models could become powerful propaganda tools due to their tendency toward censorship in certain regions.

Compiled by Sergey KostenchukPublished 2026-07-20Updated 2026-07-20

2026-07-20 Research OpenAI

🤖 AI as the Perfect Propaganda Machine

The Meta Oversight Board has warned that modern AI models (Meta, Anthropic, OpenAI) could become perfect propaganda tools. Models tend to avoid criticizing authorities in countries with strict censorship, creating an "export of censorship" effect through global AI infrastructure.

🌍 AI developers need to implement human rights audits and conduct multilingual audits to avoid copying state narratives and reducing responses to minimize legal risks.

👤 Your AI assistant may be biased due to software constraints (RLHF/Safety tuning) imposed by governments. Responses to political questions in different languages can differ radically.

Source 1: https://fortune.com/2026/07/16/claude-trump-china-ai-double-standard/

Sources