New AI Fake Victims Disrupting Scammers’ Business Models

Cybersecurity experts at Macquarie University have discovered a multilingual chat designed to keep scammers engaged in long fake phone calls and ultimately reduce the billions of people who lose money to global criminals every day. invented a bot.

A new AI-driven system hopes to put an end to this problem by creating compelling fake victims in the form of multilingual chatbots that waste the time of scam callers. I am aiming. estimated $55 billion People lose to thieves every year.

Keep talking: The new platform uses voice clones to keep scammers on alert with fake conversations with AI chatbots.

Named Apate after the Greek goddess of deceit, the system uses convincing voice clones to converse with real imposters to “fool the impostor.”

Professor Dali Kaafer, executive director of Macquarie University’s Optus Cybersecurity Hub, said: “Telephone fraud is carried out by organized crime groups and currently only a fraction of the criminals are caught and the money is unlikely to be recovered. Almost none,” he says.

The idea came to Professor Carfer while having lunch with his family. Then I got a call from a scammer. He pretended to be funny to make his kids laugh and kept the crooks nervous for his 40 minutes.

“I realized that scammers were wasting their time trying to stay away from vulnerable people. That’s what it is,” says Professor Dali.

“Then I started thinking about how to automate the whole process and develop a computerized chatbot that could use natural language processing to have a trusted conversation with the fraudster,” he says.

Kaafer says his team is currently applying for a patent on this highly effective technique.

“We are excited about the potential of this new technology to aggressively disrupt the fraud phone business model and make it unprofitable,” he says.

The lucrative global phone fraud trade is growing year by year, with the ACCC estimating that Australians lost more than $3.1 billion to scammers in 2022.

scrape off easy profits

Professor Carfer said that despite telecommunications providers blocking well over 500 million fraudulent calls since 2020, Australians are still inundated with such fraudulent calls, with only a few going through. Some said it could wreak havoc on victims.

Phone fraud is on the rise globally for several reasons, he says.

Disruptor: Cybersecurity innovators Michal Koepkowski, Ian Wood, Nardine Busta, and Professor Dali Kaafer have announced a new voice designed to stop scammers around the world from stealing money from victims. Developed a technology bot.

Technologies such as Voice over Internet Protocol (VOIP) make it easy and cheap for cybercriminals to hide their location and pretend a call is coming from any number.

On the technology side, however, updating telecommunications infrastructure and protocols to improve call authentication is difficult and expensive.

“Economically, this is a high-profit, low-cost ratio for scammers, making this a highly lucrative, relatively low-risk criminal activity. It Is difficult.”

Partnering with telecom providers is key to making this new technology truly effective.

This situation has led to an increase in scammers who specialize in exploiting human emotions and fears.

“The business model of scammers relies on making big profits from a few victims. say.

“Our model locks them in, wastes time, and reduces the number of successful scams,” he says. “We can destroy their business model and make it much harder for them to make money.”

History of the birth of APATE

A team at the Macquarie University Cyber ​​Security Hub analyzes fraudulent calls and uses machine learning techniques and natural language processing to identify typical fraud “scripts” that scammers use on victims. We started by identifying social engineering techniques.

We then trained a chatbot on a dataset of real-world fraudulent conversations, ranging from fraudulent call recordings to fraudulent email transcripts to chat logs from social media platforms, and trained the bot on real-world fraudulent conversations. We’ve made it possible to generate your own similar conversations.

Thanks to advances in natural language processing (NLP) and AI human voice cloning, Professor Carfer is able to speak fluently, adopting specific personas to keep conversations on track and deliver compelling and consistent responses It is said that it has become possible to develop an AI agent that can .

“We developed a conversational AI bot that can trick scammers into thinking they are talking to a victim, so they can spend their time fooling the bot,” Kafer said. says the professor.

These bots can be trained in any language or accent, and since phone fraud is a global problem, this technology can be deployed anywhere in the world.

live scam call trial

The team is currently piloting the chatbot with live fraud calls, redirecting victim calls to a test prototype, an “always-on honeypot” with a wide range of personas.

Please keep talking. Professor Dali Kaafer, pictured, and his cybersecurity team hope the new anti-fraud bot will keep scammers on call for up to 40 minutes and also help banks identify the latest phone scams so they can alert their customers. doing.

“We’ve been putting these ‘dirty’ numbers all over the internet, putting them into spam apps, publishing them on webpages, etc., making them more susceptible to fraudulent calls,” says Professor Carfer.

“The bots are responding very well to some tricky situations we didn’t expect to avoid, such as scammers asking for information we didn’t train the bots for. But the bot has adapted and is coming up with a very reliable response.

“Bots are continually learning how to extend calls to achieve their primary goal of keeping scammers on the phone longer.”

The current Apate bot deployment time is already averaging 5 minutes, with a goal of 40 minutes.

Fraud-fighting bots also contribute to threat intelligence, the timely information gathered about current phone scams and their targets. This helps organizations such as large banks, retailers, and government agencies to alert their customers.

Professor Carfer said the team is in talks with a number of telecommunications providers and is open to many commercial partnerships.

“Partnerships with telecom providers will be key to making this really effective,” says Professor Carfer.

“We believe this has great potential globally, redirecting many of the spam calls that providers currently block, sending scammers to the Apate bot and binding time as much as possible. If we can do that, we won’t disrupt the entire industry, it’s doable.

“I think the ultimate meta-scenario could be that the scammers themselves deploy AI, train their own fraud chatbots, and repurpose them to converse with chatbots owned by telecom providers.

“If fraud chatbots end up talking to fraud-defending chatbots instead of stealing money from real people, I think that would be a big win.”

Apate This research is funded in part by the National Intelligence Service under the National Information Security Research Grant Program.

Professor Dali Kafar is the Executive Director of the Cyber ​​Security Hub at Macquarie University. Faculty of Science and EngineeringMacquarie University.