AI Scams on LINE & Facebook in Thailand

In Thailand’s busy digital scene, a friendly chat invite or a convincing video call can now hide a sophisticated AI-powered rip-off. From bogus government badges to deepfake job pitches, scammers are stepping up their game on platforms we use daily.

Quick glance for Thai netizens

• AI-driven impersonation is increasingly targeting LINE and Facebook users.

• Bogus officials from "Cyber Crimes Investigations Service" have defrauded victims nationwide.

• Meta’s Llama Firewall and Rule of Two aim to block malicious AI in real time.

• Collaboration with ACSC and DSI has led to the takedown of over 59,000 risky pages this year.

• Victims reported combined losses exceeding ฿89 B since early 2024.

The evolving threat landscape in Thailand

While Thailand races ahead in digital payments and social media engagement, fraud rings are exploiting every new channel. Scammers armed with stolen identities and AI-synthesized voices reach out on LINE, prompt victims to switch to encrypted chats, then drain bank accounts via PromptPay. Police data show that online crime losses topped ฿60 B by mid-2025, and many never see a baht returned.

From scam factories to personal appeals

Just across the border, so-called "scam factories" in Cambodia and beyond churn out scripts that mimic Thai slang and official insignia. These networks craft deepfake video calls impersonating senior officers or bank executives, promising to recover lost funds—only to vanish once the transfer goes through. The scale is staggering: Meta removed over 6,400 Facebook assets tied to a Cambodia-based ring this year alone.

Meta’s AI weapons: Llama Firewall and Rule of Two

To counter these high-tech schemes, Meta has rolled out two key defences:

Llama Firewall filters incoming prompts for malicious patterns before an AI model can execute them, slashing prompt-injection attempts by more than 70% in tests. Meanwhile, the Rule of Two prevents any AI agent from handling untrusted input, sensitive data, and external actions all at once—forcing a human checkpoint whenever all three are demanded.

Together, these tools form an AI safety stack that not only detects deepfake ads but also stops scammers from automating their cons at scale.

How Thailand’s law enforcement fights back

Thai authorities have teamed up with Meta and international partners to strike at the heart of these operations. Under the Anti-Online Scam Centre’s Joint Disruption Week, over 59,000 high-risk pages were removed, and six prime suspects linked to a cross-border syndicate are now facing extradition proceedings.

The DSI has also pursued cases against a network that set up 29 fake companies and 2,000 bogus websites, inflicting ฿2 B in damages. Though specific DSI-Meta task forces remain under wraps, both bodies underscore a commitment to deepen collaboration and intelligence-sharing.

Tips for safeguarding your digital life

• Activate two-factor authentication on LINE, Facebook, Instagram and banking apps.

• Scrutinise unsolicited calls, even if they display familiar faces—voice cloning is now a commodity.

• Verify any government-style credentials via official hotlines before transferring funds.

• Report suspicious chats or ads to the ACSC at 1441 or submit screenshots directly through Meta’s tools.

Looking ahead: vigilance and collaboration

Thailand sits at the crossroads of innovation and illicit AI ingenuity. For every deepfake that slips through, new guardrails like behavioural detection systems and Rule of Two are deployed. Yet technology alone cannot solve this arms race. Continuous user awareness, agile law enforcement, and transparent platform policies will determine whether we stem the tide of scams—or watch another ฿110 B slip away in 2026.