DraftKings co-founder defends company's AI use before Louisiana board

DraftKings co-founder Paul Liberman defends AI use before Louisiana board

Paul Liberman walked into a license renewal hearing in Baton Rouge on Thursday and found a room that had already made up its mind about what it wanted to know.

The Louisiana Gaming Control Board had questions prepared. They were not about DraftKings' revenue figures or its market share. They were about a New York Times report, about artificial intelligence, and about whether a company that describes itself as an entertainment platform had been using its systems to identify customers most likely to respond to gambling promotions — including customers showing signs of harm.

Liberman's response, that DraftKings' users mostly lose money anyway so targeting losers makes no commercial sense, was not the kind of answer that closes a regulatory inquiry. The board voted unanimously to extend the company's Louisiana license conditionally through February rather than renew it outright. DraftKings' existing license expires in November. The extension to February 19 creates a gap that gives the attorney general's office and state police time to finish their work.

Chairman Chris Hebert was precise about what the vote meant. He said it was not a finding of wrongdoing. He also said regulators have a responsibility to examine matters that could affect the integrity of gaming operations and the protection of Louisiana citizens. Those two statements are compatible, and the board held both simultaneously.

What DraftKings must do now is substantial. A senior executive must be designated as a board liaison within seven days. Monthly reports covering customer complaints, regulatory actions, and advertising violations are due on the fifteenth of each month. Shareholder suitability findings — including at least one investor whose institutional-investor designation has not been resolved — must be submitted within thirty days. The full renewal hearing is scheduled for February 18.

The shareholder question is worth separating from the AI question, because they are different problems with different resolution paths. The suitability gap is administrative. It has a clear end state: paperwork completed or not. The AI question does not have a clear end state. Louisiana is now part of a pattern that includes Massachusetts and Michigan, regulators asking operators to explain systems that the operators themselves have not been able to explain with precision. The Massachusetts AI probe specifically turned on whether companies could demonstrate their promotional algorithms don't function as harm-targeting tools. Liberman's argument before the Louisiana board — that targeting losers is commercially illogical — addresses intent, not mechanism.

The consensus read on this situation is that DraftKings faces a temporary procedural hurdle that gets resolved in February and that the AI allegations, being unproven, will not harden into findings. That read may be right. But the multi-state coordination of these inquiries, and the consistency of the question regulators cannot get answered, suggests something structural rather than episodic. Operators have deployed systems they understand at the output level — who received a promotion, when — but not always at the design level, meaning the criteria baked into the model. Regulators are asking about the design level. The February hearing will turn on whether DraftKings has found an answer to that question, not just a restatement of commercial intent.