DraftKings faces scrutiny over AI-driven promotional targeting in the US

A New York Times investigation says DraftKings built a machine learning model in 2023 to estimate which customers were most likely to keep gambling and lose money after receiving promotions. For PSPs and other payment partners, the key point is simple: this is what targeted bonus spend can look like when customer data, analytics, and regulated gambling economics are all in the same room.

1. The model was reportedly trained on customer betting data and assessed factors including betting frequency, account balance, and historical losses. The New York Times said customers were then assigned scores estimating how much additional money they could generate through promotions.
2. According to the investigation, former DraftKings data analyst Jayden Butts, who worked on testing the model, said the company wanted to direct free bets and bonuses toward customers expected to gamble and lose more. The investigation cited interviews with more than 40 former DraftKings employees, plus internal documents, Slack messages, and betting data from tests involving customers.
3. DraftKings rejected that characterization. The company said its promotions are “directed towards customers who demonstrate sustained, engaged use of our platform, not toward customers based on their losses.”
4. Chief Responsible Gaming Officer Lori Kalani said DraftKings also considered technology to identify customers at risk of developing gambling problems, but abandoned those projects because it determined there was insufficient evidence that it was effective.
5. DraftKings has acknowledged using AI to personalize promotional spending and said data analytics increased margins on promotion-driven sports wagers by 13 per cent last year.