Tomas Kapler wrote that the company Apate has built a fleet of approximately 350,000 AI agents who impersonate trusting individuals and communicate with scammers for as long as possible. According to him, they gather information such as bank account numbers, cryptocurrency wallet addresses, phone numbers, and details about fraudulent websites and scenarios used. The company passes this data on to banks, telecommunications operators, and other institutions. Kapler reported that Apate has recorded over 2.5 million conversations with scammers, which together took about 17 years, and obtained more than 250,000 intelligence data points. He mentioned examples such as the Commonwealth Bank of Australia, which, according to him, saw publicly reported fraud losses drop by 76% after implementing the system, and TPG Telecom, which redirected nearly 590,000 suspicious calls to Apate's agents. One other Australian bank, according to his post, broke up 22 fraudulent campaigns based on the information obtained. He added that the company secured $8.15 million in a seed round in August 2026, just shy of $10 million in total, and planned to expand into London and the United States. Kapler also wrote that Apate expected the proliferation of AI among scammers would lead them to spend more time communicating with its agents instead of potential victims. Finally, Kapler asked whether a similar system would work in Czech. He noted that Czech models may not be as advanced, but older individuals might not regard any mishearing or misunderstanding as a sign of artificial intelligence. He also mentioned that fraudulent AI farms from India, China, Russia, and Ukraine might not recognize such a deception. #AI #podvodníci #technologie (translated)





