New York imposes data-sharing rule on nine licensed sportsbooks

New York cross-operator data rule puts nine licensed sportsbooks on notice

Brian O'Dwyer did not leave the industry much room. After the New York State Gaming Commission voted unanimously to advance its revised AI restrictions, the chairman told the nine licensed operators something the legislation itself does not quite say: solve the cross-platform problem yourselves, or the Commission solves it for you, and you will not like the solution.

The restriction on AI personalisation has drawn most of the attention, and that coverage is warranted. Prohibiting operators from using algorithmic profiling to tailor promotions and wagering recommendations is a significant constraint on the acquisition and retention mechanics that drive mobile sports betting revenue. But the more consequential proposal may be the one sitting beside it — the requirement that operators identify at-risk bettors and designate a person to engage them directly.

The difficulty O'Dwyer named is structural. A customer flagged and restricted by one operator can open an account with a second, a third, and a seventh. New York has nine licensed books. The responsible gambling infrastructure each operator maintains is, at present, siloed. A patron cooling-off period means nothing if it applies to one platform in a market where eight others are a download away.

O'Dwyer's public statement was essentially a negotiating position: share the data voluntarily, or we mandate it. The industry will resist, because shared flagging data raises questions about competitive intelligence, liability, and — depending on how the requirement is written — potential antitrust exposure. These are not trivial objections. They are also not arguments that tend to move state gaming regulators who are answering to a governor whose programme explicitly targets underage gambling and marketing to at-risk customers.

Governor Hochul included both proposals in her State of the State programme announced in March. The Commission's AI rule is modelled on New Jersey's approach, which restricts rather than prohibits the underlying technology. That framing matters: a restriction is easier to defend in court than a ban, and New York's operators will look closely at how New Jersey's version has been enforced before deciding how hard to push back during the additional public comment period.

The consensus read on this story is that the AI personalisation rule is the headline risk. The data-sharing mandate is treated as a softer, longer-horizon problem. That framing has it backwards. Operators have already built compliance infrastructure around content restrictions. They have not built shared patron-flagging systems, and building one would require either a Commission-run clearinghouse or an industry consortium with its own governance problems. Either path takes longer and costs more than reconfiguring a recommendation algorithm.