Artificial intelligence may offer a powerful new way to detect illegal insider trading, according to research by Dr. Solmaz Batebi, an assistant teaching professor in the University of Washington Bothell School of Business. Although the U.S. Securities and Exchange Commission prosecutes about 50 insider trading cases each year, researchers believe this represents only a small fraction of the activity occurring in financial markets.
In her recent study, Batebi found that machine‑learning methods significantly outperform traditional models in predicting both the likelihood and size of insider trades. The research also shows that AI can dramatically reduce investigation time, cutting analyses that once took months down to minutes.
Batebi’s work further identified behavioral differences between male and female insiders, with AI models showing particularly strong predictive gains for female traders.