A new data-driven machine-learning model allows AI to discern when CEOs are lying or using deceptive language.
Researchers from the business schools at Tempe-based Arizona State University, University of Nevada, Las Vegas, University of Texas-San Antonio and Boise State University in Idaho used machine-learning models to identify deceptive language used by CEOs, as well as analysts' suspicion of deception on earning calls. Their findings were published Sept. 11 in the Strategic Management Journal.
The study focused on fraudulent financial statements. So far, AI is 84% accurate when identifying business leaders' deception, according to Jonathan Bundy, PhD, associate professor in the department of management and entrepreneurship at Arizona State.
"There has been significant research on why CEOs might commit fraud in their accounting statements," Dr. Bundy said in an Oct. 20 interview for Arizona State's website. "The answer is almost always the same: to improve stock market performance. This might be to hide something or to cover for mistakes or bad performance in the past, it might be to exaggerate good performance or to project positive performance into the future or it might be to make the firm look better compared to peers."
The study also identified a tendency among financial analysts — particularly high-status ones — to favor deceptive CEOs, assigning them superior recommendations and influencing the stock market. This project was inspired by the Wells Fargo scandal in 2016, Dr. Bundy said; the researchers wanted to understand analysts' surprise despite red flags in the financial data and language being used by leadership.
"We couldn't help but wonder how helpful that pressure would have been before the scandal broke and how much damage was done simply because the people motivated to pay attention — e.g. stock market analysts — largely failed in their duties," Dr. Bundy said. "Financial analysts get paid a lot of money to scrutinize firms to determine their 'true value,' yet they completely missed signs of this massive scandal. We wanted to know why. We wanted to see if we could use advanced machine learning to detect the possibility of deception perhaps better than analysts and other market observers."
Some correlational evidence came out of the study, too. The results show that female and older CEOs engage in less deception, while higher-performing and award-winning CEOs tend to be more deceptive — but more research is needed on these differences, according to Dr. Bundy.
The algorithm used in the study is available for free here.