AI Is Getting Really Good at Messing With Cybercriminals (www.wired.com)

Australian company Apate has developed a novel AI-based system that targets phone scammers by diverting them to AI bots designed to engage and prolong conversations without falling for the scams. With around 350,000 bots, the platform aims to frustrate scammers, allowing for the collection of critical intelligence on various scams, including URLs and bank details. This approach not only slows down scam operations but also protects potential victims by keeping scammers occupied, thus reducing their capacity to target others. The significance of this development lies in its potential to revolutionize anti-scam efforts in the AI/ML community. By leveraging automation, Apate's system exemplifies how AI can undermine cybercriminals’ operations while simultaneously gathering valuable data for law enforcement and banking institutions. Additionally, the incorporation of advanced language models into cybersecurity strategies—such as “honeypots” that appear more realistic—has shown to effectively deter attackers, suggesting a promising frontier where psychological nuances in AI could be used to combat cybercrime. This approach highlights a pivotal shift towards proactive, AI-driven defenses in an era of escalating online fraud.