A revised New York AI gambling proposal would restrict personalised promotions and bet suggestions generated from player data. A commission memorandum dated 5 October describes substantive changes following industry feedback.
The proposal appears on the New York State Gaming Commission’s 9 October agenda. The published materials reviewed for this article do not establish adoption or an effective date.
Where the draft draws the line
The draft covers automated profiling as well as AI, including tools supplied by vendors. Human approval would not exempt prohibited activity. It preserves systems used solely for fraud prevention, anti-money laundering, age checks and responsible-gambling interventions that reduce wagering. A separate, conditional exception covers opt-in pari-mutuel racing advice.
Licensees would also need system inventories and seven years of records showing inputs, outputs and model versions, with vendors subject to contractual compliance requirements.
What operators could examine now
For operator and supplier teams, our assessment is that a useful starting point is to map where personal data changes a customer-facing decision. A campaign can involve several products between selecting an audience and delivering an offer. Reviewing the campaign’s wording alone could miss the selection or timing logic.
Consider a hypothetical promotion that uses identical copy for everyone, while a model chooses when each player receives it. A product review could trace the model’s inputs, the decision it produces and the system that sends the message. That would give commercial, technical and compliance colleagues a shared description to assess.
A supplier discussion could then focus on what the operator can actually inspect. Teams could request a sample decision record and check whether it explains why a particular output occurred. A record of human approval would provide only part of that account if another system subsequently changes the audience, timing or offer.
Our coverage of the separate Massachusetts sportsbook AI review raised related questions about operator oversight. New York’s draft gives product teams a specific text to examine alongside their existing workflows. Any assessment should preserve its proposed status and distinguish an internal design choice from a requirement already in force.