DraftKings faces class action lawsuit over AI use in Massachusetts

DraftKings AI scrutiny reaches Massachusetts courts as class action filed

A class-action lawsuit filed in Massachusetts alleges that DraftKings used artificial intelligence tools to identify customers most likely to respond to gambling promotions — including individuals already showing signs of gambling harm. The case is in its early stages, and none of the claims have been proven. But the filing lands in a regulatory environment that is already moving against the company on two fronts.

Louisiana's Gaming Control Board voted unanimously on October 15 to extend DraftKings' existing licence rather than renew it in full, holding the company to a conditional approval through February 19. The board's concerns turned on two separate matters: a shareholder who had not completed the required suitability process, and the AI targeting allegations surfaced by a New York Times investigation. Assistant Attorney General Delancey Houston told the board the division needed additional time to investigate. The full renewal hearing is now scheduled for February 18.

Paul Liberman, DraftKings' president of operations and co-founder, appeared before the board and argued that the AI targeting narrative was conceptually flawed — most customers lose money, he said, so the incentive to identify and exploit a losing subset collapses under its own logic. Chairman Chris Hebert acknowledged that news reports are not findings of regulatory violations. The board extended the licence anyway.

That distinction — between an allegation and a finding — is the ground DraftKings is standing on. The problem is that ground is narrowing. Michigan has joined Massachusetts in reviewing the company's AI practices. Louisiana has imposed monthly compliance reports, a senior executive liaison requirement, and a written questionnaire covering AI use, prediction market activity, and advertising conduct. The regulatory apparatus is not waiting for the litigation to resolve.

This is where I differ from the consensus framing, which treats the Louisiana extension as a procedural delay and the Massachusetts suit as a separate matter. I have watched enough multi-jurisdiction regulatory actions to know that the sequencing matters more than the individual outcomes. When regulators in three states begin asking the same question in the same quarter, the answer they eventually get becomes the industry standard — not just for DraftKings, but for every operator running AI-assisted promotional systems. Louisiana's written questionnaire, due within thirty days, will produce a documented response from DraftKings that Massachusetts and Michigan will read.

Liberman's argument to the Louisiana board is intelligent and not obviously wrong. But it addresses the targeting rationale, not the targeting capability. A system that can identify a customer's loss trajectory and responsiveness to promotions does not need a profit motive to raise a responsible gaming concern — it needs only to exist.