Tackling AI-driven crime: The prosecutors' response

They must battle technology with technology. Global cooperation also plays a key role.

- Lucien Wong

From deepfakes powered by artificial intelligence (AI) to crypto-fuelled money laundering, criminals are finding faster, smarter and harder-to-detect ways to commit crimes. This surge in AI-enabled crime poses urgent new challenges for prosecutors worldwide.

A recent report by Interpol on Human Trafficking-Fuelled Scam Centres, noted that the use of AI has been observed in a growing number of scam cases globally, many of which originate from online scam centres in South-east Asia. The report revealed that AI was used to generate convincing fake job ads to attract human trafficking victims as well as online photos or profiles through "deepfake" technology for sextortion and romance scams.

In Singapore, the Singapore Police Force's Annual Scams and Cybercrime Brief noted that there have been more than 51,500 reports of scam cases in 2024, with losses of at least $1.1 billion. The main types of scam cases encountered involved e-commerce, job, phishing and investment scams, with the volume and effectiveness of these scams likely being enhanced by AI tools.

What we are seeing in Singapore is part of a wider global shift: scams are no longer simple confidence tricks, but increasingly AI-enabled, borderless operations. AI has made crimes more sophisticated and harder to detect. It has also helped criminals overcome language barriers and transcend international borders.

From the prosecution's perspective, crime fuelled by AI poses several challenges. One, clearly, is the issue of attribution. AI-powered cyberattacks (for example, automated phishing, malware generation, deepfake voice scams) have made it harder for prosecutors to attribute actions to a specific accused person. This makes it a challenge for prosecutors to prove beyond reasonable doubt who actually ran the AI tool that committed the offence.