DraftKings AI targets losing bettors with ML, stalls addiction safeguards

DraftKings built a machine learning model in 2023 that scores each customer based on their betting history to identify which gamblers are most likely to respond to promotional offers by betting and losing more money. The higher the score, the more money the company expects to collect per promotion. The model then prioritizes those high-score users for free bets and bonuses.

Revenue model drives the targeting

The economic logic is straightforward: DraftKings makes money when bettors lose. According to former data analyst Jayden Butts, who tested the model, the company was looking for "traits and features that we can target that indicate a good investment." By that logic, he said, "the best investment would be a problem gambler." The scoring system aims to maximize expected losses per promotional dollar spent.

Gambling addiction prevention tools were deprioritized

While the company refined its methods to find and incentivize likely losers, four former employees told The New York Times that similar technology designed to predict and mitigate gambling addiction has been stalled or suppressed. This asymmetry matters: the same betting data could drive harm-reduction features, but the internal priority appears to have been placed on revenue optimization.

What builders should watch in regulated AI systems

For AI teams deploying targeting models in regulated industries, the DraftKings example raises three practical questions. First, how do you balance optimization against duty of care when the same signals maximize revenue for the company and losses for the user? Second, what governance prevents the suppression of harm-mitigation tools? Third, the company already faces legal scrutiny over its marketing practices, with a class action alleging predatory targeting of susceptible users. The reporting also highlights that AI can be used to track problematic gamblers and personalize experiences, but that same capability can become predatory without safeguards.

The DraftKings case shows how data-driven incentive design can become controversial when harm prevention tools lag behind profit optimization. For builders working on similar systems, structural safeguards and transparent governance are not optional extras; they are a prerequisite for long-term trust and regulatory compliance.