Artificial intelligence is rapidly reshaping the global scam economy, making fraudulent call centres harder to detect, cheaper to run and easier to scale, even as governments and an Australian start-up step up efforts to fight back.

A new report citing Interpol and Bloomberg investigations shows scam centres across South-East Asia are increasingly relying on cheap, widely available AI tools to industrialise fraud, targeting victims worldwide.

Interpol officials say criminals are using large language models, voice cloning and image-generation tools to replace what were once crude, error-filled scam scripts with polished, professional-sounding messages and realistic online profiles.

Previously obvious warning signs, including poor grammar, fake job ads and awkward phone calls, are disappearing.

AI now allows scammers to instantly rewrite scripts, switch languages, tailor messages to different regions and pivot tactics when authorities intervene.

Voice-cloning technology is also being used to impersonate family members, romantic partners or company representatives, adding emotional pressure that increases the likelihood victims will transfer money.

Despite government crackdowns in Cambodia, Myanmar and other parts of the region, Interpol warns scam centres are unlikely to disappear. Instead, AI is making them cheaper to operate and easier to relocate, with new operations emerging in Africa, the Middle East and the Americas.

The threat is increasingly relevant for Australian businesses and consumers as AI-generated voice calls and automated agents become harder to distinguish from legitimate contact-centre interactions.

Last year’s Qantas data breach, linked to an offshore call-centre supplier, highlighted the risks associated with complex global service networks.

The growing risk is also driving demand for defensive technology. Australian call-centre monitoring start-up Operata (pictured below) recently reached a valuation above $100 million after a new funding round.

The company’s software monitors both human and AI-driven call agents, stress-tests call-centre systems and can detect anomalies ranging from performance issues to AI-enabled scam activity.

Founder Romilly Blackburn said the rise of realistic AI-generated voice calls has created new security challenges for businesses, as the same technologies boosting efficiency are increasingly being exploited by criminals.

Estimates suggest global scam networks were already stealing tens of billions of dollars annually before the widespread adoption of AI, with figures expected to rise sharply.