A machine-learning model developed by DraftKings in 2023 assessed customers based on their expected losses after receiving a free bet or bonus, as reported by a New York Times investigation.

Notably, the same data framework that supported this model could have been utilized to pinpoint customers at risk of gambling-related harm; however, former employees informed the Times that this initiative was put on hold. This practice of covert digital profiling mirrors concerns raised by a recent surveillance app aimed at specific user groups.

The model took into account play frequency, account balances, loss-to-wager ratios, and an estimate of whether a user was likely to stop gambling altogether, according to the Times. A higher score indicated a customer was expected to generate more revenue per promotional dollar spent.

Former employees said that another project designed to create risk scores for identifying customers with potential gambling issues was abandoned. Although the infrastructure to analyze data in both directions was reportedly in place, only one model reached completion.

This trend of discreetly monitoring users without their knowledge has attracted attention across various industries. Lori Kalani, DraftKings’s chief responsible gaming officer, stated to the Times that the company actively observes customers for risky behavior. She further noted that DraftKings chose not to implement risk-prediction technology due to a lack of sufficient evidence to support it.