The Rise of AI-Driven Scams and the Challenges of Regulation
In a scam center in Myanmar, Safeer Mohammed Koorimannil impersonated a 28-year-old Singaporean woman named Ella. He targeted over 100 people daily, operating under pressure as supervisors patrolled with electric batons. In just a month, he targeted around 50,000 people across 17 countries. His victims varied widely, including a tailor in Kurdistan, a pastry chef in Turkey, and an engineer in Russia.
Koorimannil’s operation relied on AI technology from American companies. These tools were pivotal in enhancing the effectiveness and reach of scams, allowing him to operate across multiple languages and regions. He mentioned the systemic use of technology within these centers, describing how these tools enabled an industrial level of fraud.
“Everyone is a robot there,” Koorimannil stated from his home in southern India, speaking in Malayalam.
A collaborative investigation by AP and ‘FRONTLINE’ revealed how American technology supports these scams. Despite the technical potential for prevention, the legal and economic incentives for these companies remain limited. The Federal Trade Commission estimates scams cost Americans around $200 billion annually. Much of the focus has been on social media platforms, but the crucial digital infrastructure lies deeper in the supply chain.
AI and Internet Providers in Scam Networks
American-made AI models like ChatGPT and Gemini are frequently misused in scams, allowing for seamless multi-language communication and surveillance. TRM Labs’ blockchain analysis showed that scammers utilizing these models earned tens of millions of dollars. The internet infrastructure, supported by companies like Cogent Communications and AT&T, facilitates scam compound activities, particularly in Myanmar.
Starlink, Elon Musk’s satellite internet company, is identified as a major internet service provider in Myanmar, serving many scam centers despite public criticism and regulatory efforts. New satellite imagery highlights the rapid expansion of scam compounds since a crackdown along the Thai border.
The Global Response
Regulatory changes are emerging outside the United States, with countries like the UK, EU, and Singapore enforcing stricter regulations against scams. In Washington, however, the approach remains voluntary, with limited cooperation from tech giants against scams.
OpenAI and Google have programs intended to curb abuse, but the call for company accountability continues. For instance, Starlink claims to have a zero-tolerance policy for illegal activity, but usage data tells a different story, with a surge of over 2,500 units of their service in Myanmar being used for scams.
Human Impact and Legal Implications
The human toll of these scams is substantial. Victims are often coerced into losing their life savings, as seen in the case of Chris Colocousis from Massachusetts. He fell for an elaborate scam, ultimately costing him $400,000.
Advocates argue that ISPs should leverage their unique position to more proactively block fraudulent activities. Despite measures like blockchain analysis, stopping AI abuse is complex due to the dual-use nature of such technology.
This report is part of a collaboration between The Associated Press and PBS’s “FRONTLINE,” aiming to shed light on international scams and tech company responsibility.