Seldom has a corporate use of artificial intelligence sparked such immediate controversy. A New York Times investigation has revealed that DraftKings built a machine-learning model designed to identify gamblers most likely to lose money. The company then allegedly targeted those individuals with free bets and promotional bonuses. The findings have ignited a fierce debate about the ethical boundaries of data-driven marketing.
The model, developed in 2023, examined each customer's playing frequency, account balances, and loss-to-wager ratios. Workers internally referred to the resulting metric as an elasticity score. A higher score indicated that a gambler would lose more money for every promotional dollar spent. Former data analyst Jayden Butts grew uneasy, telling the Times that the ideal target was essentially a problem gambler.
What makes this revelation particularly troubling is the company's apparent double standard. Former employees disclosed that a parallel initiative to flag at-risk gamblers was shelved. A data scientist had begun building a predictive tool in 2024 to identify customers sliding toward crisis. However, DraftKings abandoned the project, claiming the approach lacked sufficient evidence.
DraftKings has disputed the characterization of its promotional strategy as predatory. The company maintains that its promotions target customers who demonstrate sustained engagement, not those who simply lose. Executives have credited AI-driven analytics with improving sportsbook margins by thirteen percent in 2025. Nevertheless, the gap between corporate rhetoric and internal practice has drawn scrutiny from regulators.
This case exemplifies a broader dilemma confronting the technology sector. Companies increasingly possess the infrastructure to both exploit and protect consumers simultaneously. Had DraftKings invested equally in harm-prevention tools, the narrative might have been entirely different. As regulators begin to examine these practices, the gambling industry may face its most consequential reckoning yet.
