The new era of digital troll farms

Real talk: the days of needing a massive team of people to run a disinformation campaign are lowkey over. What used to take an army of humans to manage—fake websites, bot accounts, and forged documents—is now being automated by AI.

Anthropic recently dropped a report showing how their model, Claude, was exploited to build a commercial manipulation platform for Malaysian politics. The operation was wild: it used about 1,000 X (formerly Twitter) accounts and a fake news site called “Malaysia Pulse” to push fabricated documents. It even scraped census data to specifically target voters by exploiting sensitive rifts in religion, race, and the monarchy.

Why it matters right now

Southeast Asia is basically a testing ground for this stuff. Because the region is super young and hyper-connected, it’s prime real estate for these tactics. Experts point out that AI isn’t necessarily inventing new tricks, but it is supercharging the process. Instead of needing 10 people to write narratives, you can just prompt an AI to do it in seconds.

Katie Harbath, who leads the consultancy Anchor Change, notes that while AI makes content creation cheaper and more personalized, it still relies on social media to spread the lies. We’ve seen this before—like when OpenAI banned accounts using AI to support Philippine President Ferdinand Marcos Jnr. Even in Indonesia’s 2024 election, AI-generated cartoon images were used to pivot a politician’s "strongman" image into something more friendly.

How to protect your feed

It’s getting harder to tell what’s real, especially when AI scrapes legitimate news and just rewrites it under fake bylines. Dr. Ben Loh from Monash University Malaysia warns that this content often just confirms the biases people already have, making it extra effective at inciting hatred—like the recent attacks against Rohingya refugees in Malaysia.

So, what can you actually do?

  • Check the source: If a "news" site looks weird or you’ve never heard of it, don't trust it.
  • Look for bias: If a post is making you feel super angry or is reinforcing exactly what you already think about a group of people, stay skeptical.
  • Stay aware: With major elections coming up in Malaysia (2027), the Philippines, and Cambodia (2028), the "flood the zone" tactic is real. If you see a sudden wave of identical posts, it's giving bot behavior. Stay sharp, fam.

Why it matters

As AI continues to learn and evolve, it has the potential to manipulate public opinion at an unprecedented scale. With upcoming elections across southeast Asia, the ability to rapidly produce personalized, deceptive content threatens to deepen political fractures and erode trust in real journalism.