AI-powered deepfakes drive surge in global business & consumer fraud

Anti-fraud professionals are warning of a rapid rise in deception powered by artificial intelligence, as new survey results highlight both the scale and sophistication of AI-driven fraud schemes affecting businesses and consumers worldwide.

Deepfake surge

A survey conducted among members of the Association of Certified Fraud Examiners (ACFE) found that 77% of respondents have seen an acceleration in deepfake social engineering incidents during the last two years. Looking ahead, 83% anticipate a further increase, with 55% expecting a significant rise in such fraudulent activity over the next twenty-four months.

The findings come as concern grows over the emergence of advanced tactics used to commit financial and data crimes, with AI enabling fraudsters to create more believable and effective schemes at greater speed and scale. John Gill, President of the ACFE, said:

"Artificial intelligence has become one of the most powerful tools in business - and one of its most potent threats. Awareness is our best defense as new risks continue to evolve. Educating professionals, equipping government and industry and empowering the public to recognise the AI-guided threats proliferating unseen is vital to maintaining trust and building confidence for what lies ahead."

Preparedness challenges

Survey data indicate that fewer than one in ten anti-fraud professionals currently feel well prepared to tackle AI-powered fraud. Stu Bradley, Senior Vice President of Risk, Fraud and Compliance Solutions at SAS, commented:

"AI is blurring the boundary between truth and imitation, with untold billions at stake. Even as AI drives seemingly limitless progress, it tests the very limits of truth itself. We must educate the public about what's at stake - and ready government and industry to face AI-charged fraud, at a time when fewer than one in 10 anti-fraud pros feel well prepared, according to our recent survey of ACFE members."

Sector responses

Organisations across banking, insurance, public services and digital identity sectors are increasing their reliance on AI-powered defences to keep pace with changing threats. In Norway, BankID, the national digital identity provider, is integrating its authentication signals into SAS' real-time fraud scoring and decisioning systems. David Sæle, Product Manager, BankID Anti-Fraud, said:

"BankID goes beyond identity protection - it powers intelligent fraud prevention. By combining our identity signals with SAS' AI-driven fraud analytics, we've moved from reacting to fraud to anticipating it. The result is smarter real-time detection and fewer false positives, enabling faster, more confident decisions that protect both users and trust at a national scale."

In the United Arab Emirates, Ajman Bank has implemented SAS' real-time fraud management platform to monitor customer behaviour and identify threats across payment channels. Abhishek Sharma, Chief Risk Officer at Ajman Bank, said:

"Our partnership with SAS and DataScience ME reflects our commitment to adopting world-class technologies that protect our customers and ensure the integrity of our banking operations. With real-time analytics and tailored models, we are delivering smarter, safer banking for our community."

Insurance and public sector

DB Insurance in South Korea has deployed network analytics to uncover and track fraudulent networks involving interconnected repair shops, clinics, and brokers. According to a senior claims operations leader at DB Insurance:

"We turned it on, and just like that, dozens of cases lit up revealing invisible fraud connections. We could see 10 million customers, every claim, every connection. We stopped reacting to fraud and started preventing its spread."

In the US, a southern state has used machine learning to improve its Supplemental Nutrition Assistance Programme (SNAP) fraud investigations and processing times. A senior official responsible for the state's program integrity oversight said:

"We saw a 50% reduction in the processing time for our investigations, and we moved from a 12-month processing time down to six months. That was huge for us, because we were also facing budget constraints and couldn't add more resources."