The Massachusetts Gaming Commission will examine how every licensed sportsbook in the state uses artificial intelligence and machine learning, beginning with DraftKings after allegations that it used the technology to identify gamblers most likely to lose.
Chairman Jordan Maynard said the immediate task was to “understand the specifics that were reported” by The New York Times. The commission’s newly formed AI task force and Executive Director Dean Serpa will interview operators before the regulator decides whether AI- and machine-learning-specific rules are required.
The review will start with DraftKings’ Massachusetts operations before extending to the other licensed online sportsbooks. It will cover the use of AI in customer acquisition, promotions and responsible gaming, rather than focusing only on marketing. Maynard said any further action or policymaking would be decided “when and if appropriate.”
The inquiry follows a New York Times investigation based on interviews with more than 40 former DraftKings employees, internal research memos, Slack messages and customer betting records. The investigation alleged that the company directed staff to build a model to identify customers who would respond to promotions and were likely to lose.
A separate account of the reporting said the model, built in 2023, rated customers on “elasticity”, or responsiveness to incentives such as free bets and other offers. Former employee Jayden Butts said the commercial question was whether a customer would give DraftKings more money than the company gave them in promotions. The reporting also said highly elastic slots players spent more than less-elastic players in early 2024.
DraftKings rejected the implication that its marketing unfairly or improperly targets customers. The company said promotions are directed at users showing sustained engagement on its platform, rather than based on their losses.
The reporting also described efforts by former employees to create models that could identify when a customer might be approaching a crisis and might need intervention. One such model was said to have shown promise, but a planned presentation in early 2025 was canceled and two comparable projects were shelved. Lori Kalani, DraftKings’ chief responsible gaming officer, said the company monitors potentially risky behavior and needs customers to bet within their means; she said evidence showed risk-modeling technology was not helpful.
Massachusetts had already commissioned research into the issue. A one-year, competitively awarded project was awarded to the UNLV International Gaming Institute in July 2024 and produced a report in November 2025 on AI, player-risk indicators and financial risk. It recommended that the commission establish an internal AI champion or task force to monitor licensees’ use of the technology.
The UNLV research identified AI applications across operational efficiency, customer relationship management, engagement, and compliance and risk, including offer optimization and anti-money-laundering detection. It warned that advanced personalization could improve engagement while creating ethical risks when behavioral or demographic data are used to target vulnerable people.
Its systematic review of 68 studies found 65 distinct behavioral indicators of gambling risk across play, engagement, profile data, responsible-gambling tool use and payments. Payment indicators had the strongest evidence, with deposit number and amount consistently among the highest-ranked measures. The report also said proprietary commercial risk-detection systems often lack sufficient methodological transparency for independent review and regulatory oversight.
