How AI could ‘supercharge’ election risks across south-east Asia

A recent threat assessment by Anthropic has highlighted the growing danger of AI-driven disinformation in south-east Asian elections. The report details a failed influence operation in Malaysia that utilized Claude AI to generate fake news, fabricate documents, and manage a network of 1,000 social media accounts to exploit political and social divisions. Experts warn that AI significantly lowers the cost and effort required to produce personalized, high-volume propaganda, effectively acting as an automated 'troll factory.' While platforms like OpenAI and Anthropic are implementing safeguards, the region remains highly vulnerable due to its hyper-connected, young populations and existing political polarization. From Indonesia to the Philippines, AI is already being used to manipulate public perception and reinforce deep-seated biases. Analysts suggest that the next frontier involves 'flooding the zone' with AI-generated content to influence the data sources used by large language models, posing a persistent challenge for democratic integrity.