Truace
Subscribe

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

Years ago it would have been an elaborate operation involving an army of cyber troops creating fake news websites, social media accounts and forged dossiers, all deployed to spread disinformation en masse. Now, all you need is AI. In a worrying sign of how AI could “supercharge” fake news, Anthropic recently revealed how Claude AI was used in a political influence operation in Malaysia. It created a commercial election manipulation platform, utilising a network of about 1,000 X accounts, a fake news outlet calle…

By the Truvace desk2 min read1 sourceJournalismTRV-2026-1502· cite & verify
How AI could ‘supercharge’ election risks across south-east Asia

In brief

Anthropic's September threat assessment described how its Claude model was used to build a commercial election manipulation platform aimed at Malaysian voters. The setup combined a network of roughly 1,000 X accounts, a fake outlet named Malaysia Pulse and fabricated documents, and drew on census and electoral data to target social fault lines across all 222 parliamentary constituencies.

The case matters because it suggests that work once needing a troll factory can now be done by prompting a model, in a region with several elections due between 2027 and 2029. Experts quoted say distribution still depends on social media platforms, and the Malaysian effort was ineffective once Anthropic intervened, so the real-world reach of such operations in future votes remains uncertain.

Main points

  1. Anthropic reported that Claude was used to create a political influence operation targeting Malaysia's recent state elections.
  2. The operation used about 1,000 X accounts and a fake news outlet called Malaysia Pulse.
  3. The fake outlet scraped Malaysian reporting, had the model rewrite it, and republished it under fabricated bylines.
  4. Katie Harbath says AI makes content more personalised, quicker and cheaper, but distribution still depends on social media platforms.
  5. Upcoming polls include Malaysia (possibly late 2027), the Philippines and Cambodia (2028), and Indonesia the following year.

The problem

An AI-built influence platform used about 1,000 X accounts and a fake news outlet to target Malaysian voters, showing how AI makes election manipulation faster and cheaper in a disinformation-prone region.

The rundown

According to Anthropic's report, the Malaysian operation pulled in census and electoral data to exploit divisions of race, religion and royalty. The fake news outlet also republished articles from Russian and Chinese state-aligned outlets, including TV BRICS, Xinhua, Sputnik/RIA and CGTN, with attribution stripped. Anthropic said it disrupted the activity, strengthened its safeguards and shared intelligence with authorities and industry partners where appropriate.

The article places this within a wider regional pattern. Last year OpenAI banned accounts using AI to generate comments about Philippine politics, many supporting President Ferdinand Marcos Jnr. In Indonesia, an AI-generated cartoon of Prabowo Subianto helped soften his strongman image and became wildly popular in the 2024 election, and AI content has also been used to inflame hatred toward Rohingya refugees in Malaysia.

Researchers quoted say the effect is partly to speed up production and to reinforce existing beliefs. Rossine Fallorina says AI mimics a troll factory and removes the need for many people, while Ben Loh says AI-generated content reinforces what viewers already believe. Harbath warns that actors could flood the zone to influence large language models, especially where news is scarce or mostly government-sponsored.

Sources

  1. JournalismThe Guardian2026-10-01

The debate