Anthropic Disrupts AI Influence Operation Targeting Malaysian Voters
Anthropic disrupted a political influence operation that used its Claude AI to target Malaysian voters through fake news and social media manipulation.
AI company Anthropic revealed in a September threat assessment that its Claude AI was used in a political influence operation targeting voters across all 222 Malaysian parliamentary constituencies. The operation employed a commercial election manipulation platform featuring approximately 1,000 X accounts and a fake news outlet named Malaysia Pulse.
Actors including governments, state-aligned propaganda institutions, state media, and private influence firms used census and electoral data to exploit social fractures regarding race, religion, and royalty. The Malaysia Pulse outlet scraped legitimate local reporting and republished articles from Russian and Chinese state-aligned outlets, such as Sputnik and Xinhua, while removing original attribution.
Anthropic stated it "disrupted the activity, used what we learned to strengthen our safeguards, and shared intelligence with authorities and industry partners, where appropriate." The incident follows similar patterns of AI-generated disinformation seen in Indonesia and the Philippines, raising concerns about the weaponization of AI in future Southeast Asian elections.
