DraftKings targeted losing bettors with AI promotions, Times reports

DraftKings developed machine learning models to analyze customer data and target betting incentives at users most likely to bet and lose more money, according to a report published by The New York Times on 19 September 2026. The investigation was based on interviews with more than 40 former employees as well as internal research memos, presentations, Slack messages, and experiment records.

Starting in 2023, former data analyst Jayden Butts tested an internal model that scored casino and sports bettors based on dozens of metrics, including frequency of play, daily balances, and past losses relative to wagers. Customers predicted to lose more per promotion were prioritized for free bets and bonuses. In September 2023, Butts tested the system on roughly 5,000 casino users before expanding the trials.

Promotions play a major role in the company's business model. Last year, DraftKings generated about $8.7 billion in gross revenue from sports and casino gamblers while distributing around $3 billion in promotions, according to data compiled by Citizens Bank.

While DraftKings refined tools to direct incentives toward users prone to losing, former employees stated that the company shelved an internal machine learning initiative designed to assign risk scores to detect problem gambling. Lori Kalani, chief responsible gaming officer at DraftKings, confirmed the company declined to deploy predictive risk technology, arguing there was no evidence demonstrating its utility.

DraftKings rejected allegations that its marketing practices are unfair or improperly targeted, stating its promotions are directed at customers showing sustained engagement rather than based on losses. The company also disputed the characterization of its promotional systems provided by former staff.