DraftKings faces a proposed class action alleging that it used AI to direct promotions towards customers likely to gamble and lose more. West Virginia resident Daniel Vest filed the case in the US District Court for the District of Massachusetts on 30 September 2026.
DraftKings denies the allegations. In a statement to WBUR, which reported the lawsuit on 1 October, spokesperson Park Winslow said the company does not use AI to select promotional recipients because of their losses or signs of potentially problematic gambling. He said it intends to defend litigation on the matter.
The dispute centres on customer data
In the complaint, Vest argues that the alleged use of personal information conflicts with responsible-play commitments in DraftKings’ privacy notices. He alleges that he received roughly 70 promotional messages across several channels during the month leading up to 25 September.
The filing relies heavily on a September New York Times investigation, which DraftKings has disputed. Its three pleaded claims concern breach of contract, breach of implied contract and the return of money allegedly retained unfairly. A footnote expressly says that a Massachusetts consumer-protection claim under Chapter 93A has not yet been asserted.
Vest asks the court to certify a class and award financial relief, as well as prevent the alleged targeting. Those requests do not establish that a class has been certified or that the company has breached its obligations.
A separate case from the regulatory review
The lawsuit follows the Massachusetts Gaming Commission’s review of sportsbook AI use, which we covered previously. The civil claim and the regulator’s examination are separate processes. WBUR reported that the commission declined to discuss the lawsuit because it does not comment on pending litigation.
The operational question behind the case
Our editorial assessment is that the case raises a practical question for CRM and data teams: can an operator reconstruct why a particular customer received an offer? An explanation would need to connect the information used by a model, its intended objective, campaign eligibility and any player-protection checks applied before delivery.
That record would also need to distinguish a model’s recommendation from the decision to send a message. A customer may receive communications through several channels, so reviewing each campaign in isolation can miss the combined experience. These are questions operators can examine within their own processes; the complaint does not establish how DraftKings’ systems actually handled them.