There has been plenty of talk in recent years about artificial intelligence - and with it, fear about its potential both for nefarious purposes and for taking jobs from humans.
But a Royal Roads University researcher argues that AI and machine learning can be vital tools for good that, paired with human expertise, promote honest accounting practices and protect against financial statement fraud by corporations. AI can serve as an intern for analysts who want to analyze large datasets more efficiently, identify patterns and anomalies quickly and enhance their decision-making processes with advanced predictive analytics.
Prof. Mark Lokanan, an RRU data scientist and machine learning specialist, recently co-authored three articles with colleagues and graduate students. These articles were published in prestigious journals ranked A* and A by the Australian Business Deans Council (ABDC), making the case for protecting public interest, investor interests and companies' futures by leveraging advanced machine learning (ML) and artificial intelligence (AI) technologies to detect and prevent fraud in financial statements and supply chain operations.
New AI and machine learning research from Prof. Mark Lokanan can protect against financial fraud when paired with human expertise.
And they could have a significant impact on both regulators protecting the public interest as well as companies protecting their bottom lines.
"Auditors are not able to capture the minuteness of frauds embedded in financial statements. Hence, we need more robust, automated machine learning models," Lokanan says. "Research shows us that auditors can only capture about four per cent of what's happening. They have volumes and volumes of data to review that's happening behind the scenes that the naked eye can't capture."
Indeed, he says there have been notable cases in recent years, from a $200-million fraud by the international food conglomerate Kraft Heinz to an alleged fraud worth an estimated $78 billion by the Chinese company Evergrande Group.
Revenue fraud, where corporations inflate their reported income to hide signs of financial distress to shareholders, competitors, employees and consumers, is common, Lokanan says. His AI models can spot such red flags for further investigation. The models use a combination of algorithms and theories that are commonly used in criminology. In this case, the model locates the financial pressures that might drive someone to commit fraud.
But, he notes, "You also need those investigative eyes that's the human part of it. AI can't do that Now we need the forensic accountants to go in there and investigate these red flags."
Lokanan says his AI and ML tools could be valuable for regulators protecting the public interest as well as companies protecting their bottom lines.
"You either need to educate your workforce or bring in more people to address these frauds. And the initiative must be there by the companies themselves to use these algorithms they're already in place in some companies."
Read more about Lokanan's recent research:
- The use of machine learning algorithms to predict financial statement fraud The British Accounting Review (ABDC ranked A*)
- Supply chain fraud prediction with machine learning and artificial intelligence International Journal of Production Research (ABDC ranked A)
- The application of machine learning to study fraud in the accounting literature Journal of Accounting Literature (ABDC Ranked A)
