The call that says your IRS account is frozen and the DM that says “hey, you look familiar” feel like separate problems. They’re not. In 2026, they’re two products from the same global scam industry, an industry that now runs call centres in Indian villages, forced-labour compounds in the Myanmar jungle, and the AI and financial infrastructure built in the US. Victims lose money. Workers lose freedom. And the tools, phone numbers, and payment rails that make it all work are often American. There’s no clean side to this story. Everyone is either a target, unknowingly complicit, or trapped inside it.
The Front Office: Indian Call Centres and the $48 Million IRS Scam
Most people know this version. A US area code flashes on your phone. A voice says your Social Security number is linked to crime. Another says your computer has a virus and needs remote access. It sounds official, urgent, and American.
Investigations in 2026 traced one network of Indian call centres that scammed more than 650 Americans out of roughly $48 million in a year. The operation wasn’t run out of a back alley. It looked like a normal BPO. In towns where jobs are scarce, young people are recruited for “customer support” roles. On day one, they get a new name, “Mike from Microsoft” or “Officer Daniels from the IRS”, and a script.
What’s new is how those scripts are made. They’re no longer written by a supervisor. They’re generated by AI language models built by US tech companies. The AI analyses thousands of successful scam calls, learns which words create panic, and produces new versions daily. “Your IP was used to download illegal material.” “We see suspicious transfers from your bank.” The goal is to short-circuit thinking and get compliance in under 90 seconds.
The infrastructure is also American. The calls come through US VoIP providers. The fake websites are hosted on US cloud servers. The money often moves through US payment apps and bank accounts before it’s converted to crypto and sent abroad. To the victim, everything looks domestic. The caller ID says Chicago. The website ends in .com. The person sounds like they’re in the next state. That’s intentional.
The Back Rooms: Trafficking and Forced Romance Scams in Myanmar
Three thousand miles east, a different kind of scam factory operates behind fences.
Take the case of a man from Kerala who answered an ad for a “data entry job in Thailand.” He flew in, was driven to the Myanmar border at night, and walked into a compound guarded by armed men. His passport was taken. He was told his family would be hurt if he tried to leave.
Inside, his job was to run up to 50,000 fake romance and investment profiles. The compound functioned like a corporation, with shifts, KPIs, and managers, except the workers were prisoners. Their targets were in 17 countries, mostly the US, Canada, UK, Australia, and Gulf nations. They posed as US soldiers, crypto traders, and doctors working overseas. The conversation started with compliments, moved to trust, and ended with a request: send money for a customs fee, invest in this platform, help with a medical emergency.
Violence enforced the quotas. Workers who didn’t meet targets were beaten or denied food. They were both victims and perpetrators. And the tools they used were identical to those in India: AI chatbots to write messages, deep-fake photos for profiles, and US-registered websites and payment links to collect money.
This is where “scam” becomes “trafficking.” Economic pressure keeps people in Indian call centres. Armed guards keep people in Myanmar compounds.
The Connective Tissue: American AI and US Infrastructure
The reason these two operations feel like one industry is that they share the same backbone.
AI as the scriptwriter: US-made large language models are being used to generate and optimise scam messages at scale. They can write an IRS threat and a love letter in the same minute, then test which version gets more replies. They can clone a voice from a 3-second clip and call a victim’s parents. The human worker just has to deliver the line and handle objections.
US digital rails: Bulk US phone numbers, domain names, cloud hosting, and fintech accounts are purchased and used to give scams legitimacy. A call from a Florida number, routed through a Delaware-registered VoIP company, feels safer to pick up. A payment link on a US payment app feels safer to click.
Global money movement: Money from a tech-support victim in Florida can pay for security at a Myanmar compound. That compound then produces more scammers to target a retiree in Texas. It’s a closed loop. Law enforcement calls it jurisdictional arbitrage. Scammers call it logistics.
Why 2026 Is Different: Scale, Speed, and Personalisation
This isn’t the clunky “Nigerian prince” email anymore. AI removed the biggest bottleneck: human effort.
Before, you needed weeks to train a new caller. Now an AI can produce 100 script variations and tell you which one works best. Before, one person could manage 10 romance conversations. Now one trafficked worker, with AI assistance, can manage thousands. The messages are personalised, timely, and emotionally intelligent.
At the same time, the physical location matters less. A script written in California can be delivered by a forced worker in Myanmar to a target in Michigan using infrastructure registered in Texas. That’s why geography no longer protects anyone.
No Innocent Geography- This is the hardest part to face. No country is only a victim or only a perpetrator.
The US provides the technology, phone numbers, and financial system that scammers rely on. India provides both legitimate call-centre labour and the recruitment pipelines that sometimes feed trafficking. Myanmar and other border regions provide territory with weak governance where criminal groups run compounds. And consumers across 17 countries provide the money.
Everyone has a reason to look away. The tech company says it just builds models. The telecom says it just sells numbers. The bank says it just processes payments. The local official says the compound isn’t in his jurisdiction. Meanwhile, the machine keeps running.
The Human Cost- Behind the numbers are people.
The IRS scam victim in Ohio loses retirement savings and then spends months freezing credit and filing reports. The romance scam victim loses money and trust. The worker from Kerala loses years of freedom and comes home with trauma.
The $48 million taken from 650 Americans and the 50,000 fake profiles run by one trafficked man are linked by code and cash. Treating them as separate crimes misses the point.
What Would It Take to Break the Chain?
This won’t be solved by warning posters alone. It requires treating it as one supply chain.
Tech accountability: AI companies need safeguards, watermarking, and audits so their models can’t be fine-tuned for fraud at scale.
Infrastructure checks: Telecoms, domain registrars, and banks need to flag bulk buyers of US numbers, sites, and accounts used for scams.
Law enforcement coordination: India needs to crack down on fraudulent call centres and the agencies that traffic workers. International pressure is needed on Myanmar compounds, including sanctions and cutting internet/power to known sites.
Public awareness: People need to know that a US number and a .com website no longer mean “safe.”
Your phone in 2026 is the last mile of a global fraud supply chain. Until we follow the code, the cables, and the cash across all three continents, we’ll keep treating symptoms while the industry grows.
The person on the other end of the call might be trying to hit a quota. Or they might be locked in a room being forced to type. Either way, the system that put them there is the same one that called you. And it only works if every part keeps moving.
References:
- Scribd https://share.google/wVRHu2s5d6eI3ijEA
- StrongestLayer https://share.google/qerLv6m3erzse6epv
- Association of Certified Fraud Examiners (ACFE) https://share.google/ySGFeM7jr0YHP2XVA
- Global Cyber Alliance https://share.google/5JvuMtCkVkT7D4JCz