AI may replace people in Southeast Asia’s scam complexes—and that could undercut the drive to stop them

AI is increasingly automating Southeast Asia’s large scam-centre industry—which the UN estimates employs hundreds of thousands of people from more than 50 countries—with generative models already drafting outreach messages and experts warning they could soon replace many human roles in “pig butchering” frauds. Automation may reduce visible human trafficking and thus weaken external political pressure and informant flows that have driven recent crackdowns in Thailand, Cambodia and Myanmar, though Interpol cautions trafficking will be reshaped rather than ended. Criminals are also exploiting fintech, stablecoins, social platforms and even satellite internet (AFP found more than 2,000 Starlink devices in Myanmar scam hubs) to scale operations, while platforms face a fraught trade-off between aggressive moderation and false positives—Meta says it has disrupted roughly 8 million Facebook/Instagram accounts and banned 6.8 million WhatsApp accounts linked to scam centres—raising compliance, reputational and regulatory risks for tech and financial firms.

Analysis

The United Nations estimates hundreds of thousands of workers from more than 50 countries are currently held in Southeast Asia’s scam centres, and researchers report large language models are already drafting outreach messages with the potential to automate most steps of “pig butchering” frauds, which build victim trust before defrauding them. Ling Li warns AI could replace subsequent scam steps, while Interpol’s Stephanie Baroud cautions that automation is more likely to reshape human trafficking than end it, preserving criminal networks’ operational adaptability.

International pressure has driven arrests in Thailand, Cambodia and Myanmar, but experts note that if automation reduces visible trafficking and kidnappings, foreign governments and NGOs may withdraw resources and informant flows that have enabled recent crackdowns, complicating law enforcement. Arrests of thousands have occurred, yet diminished human-facing harm could reduce political will and make syndicates harder to dismantle.

Criminals are exploiting fintech (stablecoins), social platforms and even satellite internet to scale operations; Meta reports it has detected and disrupted roughly 8 million Facebook/Instagram accounts and banned over 6.8 million WhatsApp accounts since early 2025, and SpaceX disabled evidence of more than 2,000 Starlink devices found in Myanmar. This dynamic raises heightened compliance, reputational and regulatory risk for social platforms, payment providers and infrastructure vendors, while traditional banks and regulated crypto exchanges face commercial incentives to purge scam activity but will likely confront tougher oversight and remediation costs.

Key Decisions for Investors

• For investors in Meta, monitor moderation-related operating costs, engagement metrics and regulatory pressure closely and consider trimming or hedging exposure if increased enforcement leads to sustained user friction or higher compliance expense
• Favor fintech and crypto platforms with demonstrable, audited KYC/AML controls and transparent remediation processes and reduce exposure to firms lacking strong compliance as regulators target payment rails used by scammers
• For satellite and infrastructure providers, demand contractual safeguards, stronger user verification and clear disclosure around misuse risk; re-evaluate exposure to vendors whose hardware can be easily co-opted
• Stress-test portfolios for increased regulatory and reputational risk in Southeast Asia and consider reducing allocations to companies with concentrated operations in Thailand, Cambodia or Myanmar unless they present clear mitigation plans