The Massachusetts Gaming Commission has asked staff to review sportsbook AI use, following reporting about DraftKings’ promotional targeting. The review covers artificial intelligence and machine learning at licensed operators. Chair Jordan Maynard announced the review at the commission’s 24 September 2026 public meeting, according to CDC Gaming.
What the sportsbook AI review covers
Executive Director Dean Serpa and staff will first engage with DraftKings to establish the facts behind the reported practices. The commission also wants to understand how other licensed operators use these technologies. Further action or policy changes will depend on what that work establishes.
CDC Gaming reports that the commission already has an AI task force, formed after research it commissioned from UNLV’s International Gaming Institute. That gives the review an existing forum for examining the technology’s use.
Allegations and DraftKings’ response
As summarised by Focus Gaming News, The New York Times alleged that DraftKings used machine learning to identify customers likely to continue gambling and lose money after receiving promotions. These are reported allegations, not conclusions reached by the commission.
Focus Gaming News also reports DraftKings’ denial. The company says it complies with Massachusetts regulations and does not use AI to target customers based on losses or use indicators of problem gambling for marketing. The announced review does not itself establish wrongdoing, impose a sanction or introduce a new law.
Analysis: connecting promotion decisions with player protection
For operator teams, the practical question is whether they can explain an individual promotional decision. A useful internal review could trace which data influenced an offer, which eligibility checks applied and whether a player-protection signal changed the outcome. Those are editorial considerations for assessing a workflow, not additional requirements announced by the MGC.
Testing the connection between marketing and safer-gambling systems would make that assessment more concrete. Teams could examine sample decisions, document who can stop an offer and check whether staff can reconstruct what happened afterwards. A model description alone would leave much of that operational process unexplained.
The issue also connects with the governance questions in the Malta Gaming Authority’s voluntary AI Gaming Charter. The contexts are distinct: Malta’s charter provides guidance, while Massachusetts is reviewing actual operator practices. Both offer reasons for technology and compliance teams to examine AI decisions together.