Even as artificial intelligence (AI) powers up cybercrime, it’s also being deployed to thwart scammers. The Australian company Apate, named after the Greek goddess of deception, has been developing AI bots to engage phone scammers, prolonging their calls and collecting valuable intelligence. The bots, with their diverse personalities, can keep scammers on the line for over two hours, potentially saving thousands from falling victim to a scam.

These AI bots are not just answering phone calls; they’re also infiltrating scam chat groups and responding to texts. Apate’s platform, used by banks and supported by telecom companies, has collected over 250,000 pieces of information from fraudsters, including scam URLs and bank details. The system is designed to frustrate scammers, making their efforts futile while gathering critical intelligence.

But how convincing are these AI bots? Kernel Panic tested the tool and found the conversations flowed naturally, even with both parties acting as AI. The system’s success lies in its ability to appear human, with some bots even hanging up on scammers and refusing to invest in questionable cryptocurrency opportunities.

While current efforts have done crucial work, the expansion of digital scamming continues. However, AI could revolutionize anti-scam efforts, as seen in the use of honeypots. Researchers at ETH Zurich have found that LLM-powered honeypots keep attackers engaged for longer, making it more difficult for them to identify and exploit real systems. This could be a significant step forward in the ongoing battle against cybercrime.