The enhancement was named Apate, after the ancient Greek goddess of deception. It is designed to reduce the number of people criminals have time to look for.

Project details

Every year, people around the world transfer an estimated $55 billion to phone scammers. Organized crime groups are often involved in phone scams, but only a small fraction of them are caught.

The idea for the new chatbot came to Professor Dali Kaafar of Macquarie University’s Cyber ​​Security Center while having lunch with his family when he received a phone call from a scammer. The professor pretended to be a fool, made his children laugh, and kept the scammer on the line for 40 minutes.

Even though I wasted their time trying to prevent scammers from reaching out to vulnerable people, I realized that there were 40 minutes of my own life that I could never get back. Then I started thinking about how to automate this process. We are inspired by the fact that our new technology can disrupt the scam search business model and make it unprofitable.
– says the professor.

Even though telecom service providers block millions of scam calls every year, that doesn’t really save the day. While modern technology makes it easy for criminals to hide their location, change their number, and more, it is difficult and expensive to upgrade telecommunications infrastructure and protocols to improve call authentication.

Kaafar’s method is based on the fact that the scammers’ business model is based on making big profits from a small number of victims. This means that only a small percentage of the thousands of searches they make are successful. Therefore, the new AI is designed to spend as much time as possible reducing the number of successful startups.

How does it work

The developers trained the AI ​​on a dataset of real fake conversations. This includes phone calls, emails, and even social media chats. The bot can be taught any language or accent.

We’ve put our numbers all over the internet, added to spam apps, posted on websites, etc. to increase the likelihood of receiving scam calls. We’ve found that bots respond quite well to some tough situations where we didn’t expect to come out dry: when scammers ask for information we didn’t teach the bots, but the bots adapt and give very convincing answers. Bots are constantly learning to achieve their main goal: keeping scammers even longer,
– Says Professor Kaafar.

Technology has great potential. If we can divert many of the spam calls that are currently blocked by ISPs to Apate bots, keeping the scammers on the line as long as possible, the entire industry can no longer stand.