Apate's 350,000 AI bots tie up phone scammers for hours

Governments have consistently struggled to crack down on cybercriminals operating beyond their borders, which has pushed both companies and researchers toward automation as an alternative. Apate is one of the clearest examples of that shift: after two years of development, its bots now work across phone calls, chat groups, and text messages, and the company says it sells the service to banks with telecom support. The bots' value is twofold — they buy time that scammers would otherwise spend on real people, and they generate data on scam URLs, money mule accounts, and bank details that can feed into broader intelligence sharing. The same logic shows up in honeypots, the false systems long used to attract hackers. Mark Vero, a doctoral researcher at ETH Zurich, says open source honeypot providers have increasingly added LLMs to make their traps look real. In research with colleagues, he found an LLM-powered honeypot held AI attackers' attention significantly longer than a more predictable one, and the attackers flagged it as fake far less often. The body cautions that even with Apate's bot swarm, cybercriminals initiate billions of messages and calls each year, and the most sophisticated operations run industrial-scale scamming sites — so AI appears to be adding a new defensive tool rather than closing the problem.