DraftKings AI scrutiny spreads as Louisiana joins ongoing review

Paul Liberman flew to Baton Rouge to answer questions his company had not fully prepared for. DraftKings' president of operations and cofounder sat before the Louisiana Gaming Control Board on October 15 and found that a license renewal hearing had become something else: a structured interrogation about whether the company's artificial intelligence tools were being used to find bettors at their most vulnerable and push promotions toward them.

The board voted unanimously to extend DraftKings' Louisiana sports betting license on a conditional basis through February 19, with a full renewal hearing set for February 18. The license would otherwise have expired November 10. Chairman Chris Hebert was careful to note that a news report is not a regulatory finding. He was equally careful to note that regulators have an obligation to look.

What triggered this was a New York Times investigation into DraftKings' AI systems and its treatment of VIP customers showing signs of gambling harm. Liberman's answer at the hearing — that DraftKings is an entertainment platform where most customers lose money anyway, so targeting losers with promotions makes no commercial sense — did not resolve the board's concern. It reframed the question. If the system is not designed to find losing bettors, what is it designed to find, and how does the company know the difference?

Louisiana is now the second jurisdiction to put that question in writing. Massachusetts opened its own AI review earlier, and Michigan has since joined. The class-action suit filed in Massachusetts — alleging that DraftKings used AI to identify customers likely to respond to gambling promotions, including those showing signs of problem gambling — remains unproven and in early stages. But three regulators conducting overlapping investigations into the same set of AI practices is a coordination problem that Liberman cannot answer jurisdiction by jurisdiction much longer.

There is also a shareholder matter. The Louisiana attorney general's office told the board that at least one shareholder holding more than five percent of DraftKings had not completed the required suitability process. One shareholder had requested institutional investor status, and regulators had not yet determined whether that designation applied. DraftKings has thirty days to resolve the outstanding suitability findings.

I have watched companies manage multi-state regulatory exposure before. The standard playbook is to treat each jurisdiction as a separate problem, give each board enough to move forward, and let the investigations run on parallel tracks that never quite converge. That playbook depends on regulators not comparing notes. Louisiana's conditional approval requires monthly reports — customer complaints, advertising violations, changes to compliance systems — filed by the fifteenth of each month. That is a reporting structure designed for an ongoing relationship, not a one-time hearing.

The consensus read on DraftKings' regulatory situation is that Louisiana is noise and the Massachusetts class action is a speculative plaintiff's case. I think that framing misses where the exposure actually sits. The risk is not a single adverse finding in one state. It is that monthly reporting requirements across multiple jurisdictions create a paper trail that regulators in other states can eventually request. DraftKings is not managing three separate investigations. It is, slowly and without having chosen this, building a centralized record of its AI compliance posture. What that record shows will matter more than any individual hearing.