New York targets personalised gambling incentives
New York regulators have moved toward one of the broadest restrictions on AI-driven gambling marketing in the US. On 9 October 2026, the New York State Gaming Commission voted to propose rules that would ban AI-personalised gambling promotions and wager recommendations across state licensees.
The proposal is not limited to mobile sports betting. It would apply to casino operators, mobile sportsbooks, video-lottery gaming agents, racetracks, off-track betting corporations, account-wagering providers and lottery couriers. Affiliates, agents and vendors acting for those licensees would also be covered.
For players comparing casinos, the central issue is direct. New York is considering whether licensed gambling businesses should be allowed to use personal data, automated profiling and machine-learning tools to decide which offers a player sees, when the offer appears and how much value it carries.
If adopted, the rule would make transparent, rule-based offers a more important comparison point between licensed operators. It would also narrow the gap between what players see on a public promotions page and what they receive after an operator has analysed their play.
What proposed Part 5404 would ban
The proposed regulation, Part 5404, would prohibit automated decision systems, algorithmic profiling, machine learning and generative systems from using patron data to select, customise, time or deliver gambling promotions.
The covered promotions are broad. They include personalised bonuses, odds, limits, messages and product recommendations. The restriction would apply regardless of the operator’s stated purpose and regardless of whether the outcome is favourable or unfavourable to the player.
That matters because gambling offers are not only deposit matches or free spins. In practice, a promotion can be shaped through several variables: the size of the bonus, the games it applies to, the wagering requirements, the deadline, the push notification timing and the order in which products are displayed inside an account.
The proposal would also ban AI-generated or algorithmically tailored suggestions for a wager or stake. That would affect recommendations that direct a player toward a particular bet amount, market, casino game or wager type based on previous behaviour.
New York’s wording would capture both visible personalisation and less obvious targeting. An operator would not need to tell a player that AI was used. If data-driven systems shaped the recipient, content, timing or amount of an offer, the rule could apply.
The presumption puts pressure on operators
A significant feature is the proposed presumption of unlawfulness. An offer would be presumed unlawful if its recipient, content, timing or amount varied between customers in a way associated with their personal data.
That approach shifts the practical burden toward the licensee. A casino or sportsbook would need to rebut the presumption with clear and convincing evidence. It would also need logs showing that the offer resulted solely from published, objective criteria.
For players, this could make promotions easier to assess. Instead of different users receiving opaque offers based on activity patterns, operators would have stronger incentives to publish clear rules. Examples could include a stated offer for all new customers, a loyalty-tier benefit available under listed criteria or a game-specific promotion with fixed eligibility conditions.
The proposal does not ban every targeted communication. It is aimed at automated decision systems and the use of patron data to personalise gambling incentives and recommendations. A licensee may still be able to run segmented campaigns if those campaigns are based only on objective, published criteria and supported by records.
The recordkeeping point is important. A public rule is more useful to players when it can be audited. If adopted, Part 5404 would make internal logs part of the compliance framework, not just marketing administration.
Why casino comparison would change
Casino comparison often focuses on welcome offers, game range, payment speed, licensing and support. New York’s proposal adds another dimension: whether an operator’s offers are objective or individually engineered.
Personalised offers can look attractive. A player may receive a bonus matched to recent deposits, a time-limited message after a losing session or a recommendation tied to games already played. The concern is that those offers may also be timed and framed to increase continued gambling when a player is most responsive.
For New York players, the proposed rule would reduce that type of one-to-one optimisation by licensed operators. A promotion that appears because a system has detected a player’s preferred game, average stake, response time or recent losses would face regulatory risk.
That does not mean promotions would disappear. Operators could still compete through published bonuses, loyalty schemes and publicly available campaigns. The difference is that these offers would need to be less dependent on hidden behavioural scoring.
For comparison purposes, the strongest operators under such a framework would be those that explain eligibility, bonus value, deadlines and restrictions clearly. A large headline bonus with unclear selection criteria would be less persuasive than a smaller offer with fixed terms and consistent availability.
Sports betting provision points to wider player-safety agenda
The same 9 October 2026 action also included a separate proposal for New York mobile sportsbooks. That proposal would require sportsbooks to identify and contact at-risk bettors and, where necessary, close their accounts.
This is separate from Part 5404, but the two proposals point in the same direction. New York is not only looking at how operators advertise gambling. It is also examining how they respond when betting behaviour suggests potential harm.
The contrast is notable. AI-style systems would be restricted for promotions and wager recommendations, but data analysis may be expected for identifying risk. For players, the distinction is between systems used to intensify gambling activity and systems used to reduce harm.
Mobile sportsbooks would be the direct subject of the at-risk bettor proposal. The AI-personalisation ban would be wider and would reach casinos and other gambling licensees. Together, the proposals would create a stricter operating environment than a rule aimed only at sports betting adverts.
Players comparing New York-licensed brands should watch how operators describe responsible gambling tools, account reviews and promotional practices. The relevant question is not only whether a site has safer-gambling pages, but whether its incentive structure avoids personalised pressure.
What happens next
The vote on 9 October 2026 started the proposal stage. The next milestone is the New York State Gaming Commission’s rulemaking process, including publication, public review and potential adoption of proposed Part 5404. No confirmed date has been identified for the next procedural step or final decision.
Until adoption, the restrictions are not final law. Operators may argue that some personalisation helps users find relevant products or receive suitable offers. Player-safety advocates may argue that gambling promotions are different from ordinary retail recommendations because they can influence stake size, session length and loss exposure.
The final text will matter. Definitions of patron data, automated decision systems and published objective criteria will determine how strict the rule is in practice. Enforcement standards will also shape whether the regulation changes day-to-day offers or mainly affects the most aggressive targeting models.
For now, the proposal gives players a clear issue to track. A New York-licensed casino or sportsbook may soon need to show that promotions are based on transparent criteria rather than private behavioural profiles. In a market where bonuses can be difficult to compare, that would be a material change.
