As fraud schemes become more sophisticated, so do fraud investigations and detection techniques. One of the hottest new fraud-detection tools is the use of artificial intelligence to complement investigators’ efforts.
FLEXIBLE AND ADAPTABLE
Artificial intelligence refers to the use of computer systems to perform tasks that typically require human intelligence. It can involve visual perception, speech recognition, decision making and language translation.
Artificial intelligence applications are particularly well suited for fraud detection. They’re flexible and easily adaptable, meaning they can adapt to new business conditions. The applications also can find relationships that were previously unknown. And because they learn from experience, artificial intelligence applications don’t need to be programmed for all of the operating conditions under which they must perform. This is especially important in fraud detection, because not every condition can be known.
Such applications are faster and more accurate than human investigators — an advantage in a world that abounds with massive amounts of data previously unavailable. With artificial intelligence lending a hand, fraud investigators can concentrate their efforts on components requiring human input, such as one-on-one interviews and evidence analysis.
DETECTION TECHNIQUES
Several types of artificial intelligence have been used to detect fraud:
Neural networks. Neural nets have been used to detect fraud in banking and credit card transactions. A network is “trained” to identify fraudulent activity by comparing aspects such as the time, frequency, size and type of transaction with an existing model established for each customer. It sends up a red flag when it spots irregular spending behaviors.
Genetic algorithms. Genetic algorithms are search techniques based on the process of natural evolution, including such concepts as inheritance, mutation and selection. Taking the concept of survival of the fittest, they “mate” the most effective solutions to produce even better ones. In the fraud arena, this leads to the development of improved detection techniques over time. At least one bank has used genetic algorithms to develop stronger techniques for signature recognition.
Fuzzy logic. Fuzzy logic deals with figures or values that are approximate, rather than specific. It permits the consideration of grey areas, as opposed to applying only hard and fast rules or thresholds. For example, traditional fraud detection software might flag a transaction of $500 as large, and therefore suspicious, but ignore a $499.99 transaction. Fuzzy logic would also regard as suspicious some transactions just under the $500 threshold.
COMPLEMENT, NOT REPLACE
Like any fraud detection technique, artificial intelligence generally indicates only data that might be suspicious. It still takes a trained fraud expert to analyze the data, conduct an investigation and confirm that malfeasance has occurred.
This publication is distributed with the understanding that the author, publisher and distributor are not rendering legal, accounting or other professional advice or opinions on specific facts or matters, and, accordingly, assume no liability whatsoever in connection with its use.
This article was written and published by Dennis Frankeberger, CPA/CFF, CFE 909-597-1100.
The Partners of Frankeberger Vausher + Company, CPAs and Litigation Consultants, have in excess of 35 years professional litigation and expert witness experience. We consult with clients, their attorneys and or accountants on the matters listed above in support of confrontational issues requiring settlement and or potential equitable adjustments. We have the technical expertise to analyze complex situations, assist with discovery, and render independent, professional opinions.
Frankeberger Vausher + Company includes CPAs, Forensic Accountants, Certified Fraud Examiners, and includes an expert with a Master’s Degree in Taxation. Dennis Frankeberger – Managing Partner, is also the Chairman of the Board of Advisors to The Leventhal School of Accounting at the University of Southern California. He has lectured extensively regarding matters of Internal Control, Discovery, Fraud, Ethics and Taxation.
- Litigation and Arbitration – Wrongful Employment Termination
- Lost Profit Analysis – Business Damages
- Mergers & Acquisitions – Lost Earnings
- Business Valuations – Due Diligence
- Contract Disputes – Wrongful Franchise Termination
- Forensic Accounting – Fraud & Embezzlement Investigation
- Family Law – Expert Witness Testimony
For more information, please contact Dennis Frankeberger, CPA/CFF, CFE 909-597-1100
Email address: FV@FVCPAs.com
Website: www.FVCPA.com
