Anthropic recently discovered that Claude AI was used in a political influence operation in Malaysia, creating approximately 1,000 fake social media accounts, a fictitious news outlet and forged documents designed to manipulate voters in state elections. The operation targeted sensitive political divisions across all 222 Malaysian parliamentary constituencies but proved ineffective once detected, yet it has sparked alarm about how artificial intelligence could weaponise disinformation campaigns in south-east Asian elections. The case exemplifies how AI can "supercharge" fake news by enabling actors to produce personalised content faster and more cheaply than traditional disinformation tactics.

South-east Asia represents a particularly vulnerable target, with young, hyper-connected populations in nations facing upcoming elections: Malaysia in late 2027, the Philippines and Cambodia in 2028, and Indonesia in 2029. The region has a documented history of disinformation campaigns; the Philippines has been likened to a "petri dish" for weaponised fake news, whilst Indonesia saw AI-generated imagery successfully soften a political candidate's public image during the 2024 election. Experts warn that AI is essentially automating what were previously labour-intensive "troll factories," enabling governments, state-aligned institutions and private firms to deploy coordinated influence operations at unprecedented scale and speed across multiple platforms.