AI for fraud detection to triple by 2021

While only 13 per cent of organisations use artificial intelligence (AI) and machine learning to detect and deter fraud, another 25 per cent plan to adopt such technologies in the next year or two.

This is a nearly 200 per cent increase, according to a new global survey by the Association of Certified Fraud Examiners (ACFE), held in conjunction with SAS.

The inaugural Anti-Fraud Technology Benchmarking Report examines data provided by more than 1000 ACFE members about their employer organisations’ use of technology to fight fraud.

The report finds, among others, the rise in the use of biometrics (cited by 26 per cent of respondents, or about one in four organisations) as part of anti-fraud programmes.

More than half of organisations (55 per cent) also plan to increase their anti-fraud tech budgets over the next two years. 

During this period, nearly nearly three-quarters of organisations (72 per cent) are projected to use automated monitoring, exception reporting and anomaly detection. 

Similarly, about half of organisations anticipate employing predictive analytics/modeling (52 per cent; up from 30 per cent) and data visualisation (47 per cent; up from the current 35 per cent). 

“As criminals find new ways to exploit technology to commit schemes and target victims, anti-fraud professionals must likewise adopt more advanced technologies to stop them,” says Bruce Dorris, chief executive of ACFE.

“The dramatic rise of AI, machine learning and predictive modeling reveals that, beyond the hype, advanced analytics is helping investigators keep steps ahead of increasingly sophisticated fraudsters,” says James Ruotolo, senior director of products and marketing for fraud and security intelligence at SAS. 

Biometrics (such as fingerprint, facial, or keystroke recognition) is the most common emerging technology respondents said they will incorporate in their anti-fraud programmes.

More than a quarter of respondents are already using these techniques and another 16 per cent are expecting to employ biometrics within the next two years. 

Both blockchain/distributed ledger technology and robotics, including robotic process automation, are currently used less frequently than biometrics (9 per cent of  organisations for both categories), but a similar proportion of organisations expect to implement these technologies in the next two years.

The technology least likely to be adopted as part of anti-fraud programmes is virtual or augmented reality.

The survey finds only 6 per cent of organisations currently use this technology, and nearly two-thirds of the  organisations in the survey do not expect to employ virtual or augmented reality as part of their anti-fraud initiatives. 

The survey notes some industries and groups have established data-sharing consortiums, in which they feed certain data into an aggregated database that all member organizations can access. 

These initiatives are designed to help member organisations benefit from the collective data of the consortium in order to identify trends and protect themselves from known threats.

In the study, 29 per cent of respondents currently contribute to such a consortium, and another 21 per cent would be willing to contribute in the future. 

However, half of the respondents indicated that they do not have any plans to participate in a data-sharing consortium to help prevent and detect fraud.

The report notes sharing data in this way can provide large benefits, but organisations might opt out of such initiatives, due to privacy concerns and logistical challenges in disclosing their data to other organisations.

Sign up for  CIO newsletters for regular updates on CIO news, career tips, views and events. Follow CIO New Zealand on Twitter:@cio_nz

Send news tips and comments to divina_paredes@idg.co.nz @divinap