Scamming the scammers: AI chatbots can help fight against phone scams

Macquarie University cyber security experts have created a multilingual chatbot designed to keep scammers on long fake calls

Advancements in technology have paved the way for innovative solutions in combating phone scams. The executive director of Macquarie University’s cyber security hub, Dali Kaafar and his team have developed conversational artificial intelligence (AI) bots capable of engaging scammers and ultimately reducing the number of people who lose money to global criminals every day.

The technology is named ‘Apate,’ after the Greek goddess of deception and has already been patented by the team.

The team is also actively engaged in discussions with various telecommunications providers, exploring opportunities for commercial partnerships.

These bots are able to mimic human speech fluently, adopt specific personas, and maintain consistent responses throughout a conversation. As a result, scammers waste valuable time attempting to defraud the bots instead of real individuals. The flexibility of training these bots in different languages and accents allows for their deployment on a global scale, addressing the worldwide challenge of phone scams.

The research team is currently conducting live trials by employing the chatbots to intercept scam calls and redirect them to their testing prototype, an "always-on honeypot" featuring various personas. To increase the likelihood of receiving scam calls, the team has spread these honeypot numbers across the internet, including spam apps and webpages.

The bots have proven to handle unexpected scenarios adeptly, responding convincingly to scammers' inquiries that were not part of their original training. Continual adaptation and learning enable the bots to prolong conversations, ultimately achieving their primary goal of keeping scammers on the line. Presently, the deployed Apate bots have an average call duration of five minutes, with the aim of reaching 40 minutes.

In addition to impeding scams, these AI bots contribute to threat intelligence. They can gather timely information on current phone scams and their intended victims, which may be able to assist major banks, retailers, and government bodies in alerting customers to potential threats.

"If scam chatbots end up conversing with scam-defending chatbots instead of stealing money from real people – I'd consider that a significant victory!" said Kaafar, emphasising the potential transformative power of this approach in tackling phone scams.