Chatty AI bots put to work in Australia s fight against fraud

Scammers targeting people in Australia may soon find themselves talking not to victims but to bots instead.

On June 27, Australia’s largest bank Commonwealth Bank of Australia deployed 10,000 artificial intelligence bots to disrupt scam operations, engage scammers in extended conversations, and gather valuable intelligence in real time.

When a scammer calls or texts, the multilingual AI chatbots create voice clones that keep fraudsters on lengthy phone calls.

During the calls, the bots also extract as much intelligence as possible and feed real-time insights directly into the bank’s scam-control systems.

“Scammers are increasingly using AI to target Australians – we’re turning the tables by using AI to fight back. Every minute a scammer is engaging with a bot is a minute they’re not targeting an Australian,” the bank’s group fraud general manager James Roberts said.

Developed in partnership with cyber-intelligence firm Apate.ai, the initiative is based on a honeypot strategy, which refers to a security mechanism that creates a virtual trap to lure attackers.

Professor Dali Kaafar, founder and chief executive of Apate.ai, said: “In the fight against scams, timing is everything. Our intelligence gives organisations like Commonwealth Bank the edge, not just to detect scams, but to anticipate and block it before it reaches customers.”

According to data from Australia’s National Anti-Scam Centre, phone scams were the most financially damaging form of scams in 2024, with 2,179 victims losing A$107.2 million (S$89.3 million).

Text messages were the second most common contact method used by scammers, with investment scams responsible for the highest losses via this channel.

Beyond immediate engagement, the bots also help identify emerging scam trends, improve the bank’s scam detection capabilities, and support collaborative efforts to shut down scam networks industry-wide.

In Singapore, the AI model Meralion screens phone calls and intercepts potentially fraudulent ones, preventing scammers