In July 2026, the Associated Press and PBS Frontline jointly published an investigation months in the making, revealing how the global fraud industry has been systematically abusing the technology and infrastructure of American tech companies. The reporting draws on tens of thousands of leaked documents, analysis of more than 200,000 device connections, and interviews with 58 victims across 19 countries along with dozens of scammers.
The core finding is unsettling: American technology is not being passively exploited — it runs through the entire digital supply chain of the fraud business.
AI Becomes the Fraud Assembly Line
Inside Myanmar's scam compounds, trafficked scammers carry out industrial-scale operations on AI-powered software platforms built on American models — "Kongtian Smart Customer Acquisition" (KT) and "Global Social Traffic Navigator" (007TG). These platforms plug in models such as OpenAI's ChatGPT and Google's Gemini, enabling real-time translation across more than 100 languages, automated replies, and role-playing chatbots.
This means a single scammer can operate dozens of fabricated identities at once, chatting with dozens — even hundreds — of targets simultaneously.
Safeer Mohammed Koorimannil, an Indian man trafficked into Myanmar, was made to target roughly 50,000 people across 17 countries in a single month — forced to play dozens of characters at once, among them "Ella," a 28-year-old Singaporean woman. The software also monitors the scammers' performance; those who miss their quotas are beaten.
Blockchain analysis shows that a single cryptocurrency wallet used by 007TG received $860,000 in payments between April 2024 and December 2025, while the fraud networks using these tools have profited by at least $75 million.
AI is driving fraud toward full automation — in the future, human scammers may no longer be needed at all.
Starlink Becomes the Compounds' Network of Choice
SpaceX's Starlink, though it has never officially sold service in Myanmar, has become the country's largest internet service provider, with a market share approaching 20 percent. The investigation found that of the new scam compounds built after the crackdown in autumn 2025, at least 13 were using Starlink to get online.
In October 2025, Starlink cut off service to more than 2,500 terminals near the compounds, and user numbers plunged — but within just two months, usage had rebounded. Analysts point out that Starlink has the ability to sever service; yet because the fraud networks contribute so heavily to its user base, it has little commercial incentive to do so.
The Unwitting Accomplices — U.S. ISPs
In a sample analysis of device connections at four scam compounds in Myanmar, one-fifth of the signals were carried by U.S.-registered companies — Cogent Communications, AT&T, Oracle, and DigitalOcean among them. These companies are not directly involved in the fraud, but they lack the legal or commercial incentives to proactively monitor for and block abuse.
The investigation documents heartbreaking individual cases: a victim named Chris Colocousis, guided by "Eliza," lost $400,000 in retirement savings. The scammers used deepfake technology on video calls and even sent people to his door to collect cash. Colocousis believes U.S. tech companies should be doing more to protect their users.
At Dike Park, an Ethiopian engineer named Ebisa, forced to work there, was beaten so badly for missing his quota that he lost the sight in one eye; a Nigerian man, Obinna Okeadu, died after being beaten.
The U.S. Federal Trade Commission estimates that Americans lost nearly $200 billion to fraud in 2024. Globally, the figure can only be larger.
The Institutional Roots of the Regulatory Vacuum
Cybersecurity experts' verdict is blunt: U.S. tech companies lack the legal, regulatory, or commercial incentives to proactively prevent fraud — "the cost of abetting fraud is zero."
By contrast, the United Kingdom, the European Union, Australia, and Singapore have all passed regulations requiring companies to actively prevent fraud, on pain of fines. The United States has no comparable federal law.
Jeanine Pirro, the U.S. Attorney for the District of Columbia, has convened a "Scam Compound Strike Force" in cooperation with Meta, SpaceX, Google, and others — a four-day operation in May 2026 disrupted more than 1.4 million accounts and malicious traffic flows. But this reads more like a symbolic mop-up than an institutional transformation. OpenAI and Google claim to run proactive anti-abuse programs, yet after receiving the AP's findings, OpenAI banned just three abusive accounts.
From Tech Optimism to Supply-Chain Responsibility
The most valuable thing about this investigation is not that it exposes the cruelty of fraud — we have heard enough of that — but that it maps the fraud's complete technological supply chain. From AI models to satellite internet, from ISPs to cryptocurrency, every link is underpinned by the products and services of American tech companies.
This is not a simple story of "bad people doing bad things with tools built by good people." When Starlink knows its terminals are being used in scam compounds but has no commercial incentive to cut them off; when AI companies know their models are being used to generate multilingual scam conversations but respond only with vague claims that they run "proactive anti-abuse programs" — the doctrine of technological neutrality cannot stand here.
The real question is not whether AI will replace human jobs, but that AI is already powering some of humanity's cruelest exploitation. And in this system, every company that supplies the infrastructure is a knowing participant.