Massachusetts Gaming Commission Asks For AI Answers in Letter to Operators

• As expected, the Massachusetts Gaming Commission has sent a letter to each licensed sports betting operator asking for more information on AI uses
• MGC Executive Director Dean Serpa is requesting answers to several questions by Tuesday, Nov. 3
• Massachusetts regulators want more information on how operators use artificial intelligence in marketing, promotions, and customer engagement

---

The Massachusetts Gaming Commission has sent a letter to each of its licensed Massachusetts sports betting operators requesting answers to questions regarding the use of artificial intelligence and how the technology is utilized in marketing, promotions, customer engagement, and assessment of problematic or addictive gambling behavior in its customer base.

Executive Director Dean Serpa issued the letter to each licensed Commonwealth sports betting operator on Oct. 9, with a deadline for answers by Tuesday, Nov. 3.

Last month, MGC Chairman Jordan Maynard announced the commission would begin evaluating each licensed Massachusetts sports betting operators’ use of artificial intelligence and how it affects its Massachusetts customer base. The decision to do so was based on a New York Times report in which it investigated DraftKings’ use of artificial intelligence and machine learning, alleging the company used the technology to intentionally target gamblers likeliest to lose.

More Answers Needed on AI Use

Each of the Commonwealth’s licensed sports betting operators received a copy of the letter. Serpa and the MGC requested specific written responses to questions regarding the use of artificial intelligence and machine learning and how it relates to several aspects of their business.

Serpa noted in last week’s MGC meeting that its sports wagering division already started reviewing each licensed sports betting operators’ privacy policies, required disclosures related to the use of machine learning, as well as its internal controls related to artificial intelligence.

The letter is separated into eight sections, each requesting specific information on an operator’s use of AI and machine learning for the following topics:

• General use of AI and machine learning
• Data used by the system’s models, algorithms, or tools
• How the system decides who receives marketing materials
• How the system determines what promotion a patron receives
• How the system determine who and how often a patron receives marketing or promotions
• Behavioral targeting and optimization
• Responsible gaming/potential harm
• Decision-making and human oversight

Included in the letter is a question directly asking each operator if its systems can “trigger a promotion or marketing communication based on a patron’s recent betting activity, losses, deposits, or other account activity?” This question likely stems from The New York Times report in which several ex-DraftKings employees noted the company used artificial intelligence and machine learning systems to identify customers who would be most likely to respond to gaming promotions and most likely to lose.

The report included interviews with more than 40 former DraftKings employees, internal research memos, Slack messages, and betting records from customers.

The letter also asks each operator to if it uses a patron’s increased losses, increased frequency of wagering, or other potentially harmful gambling behaviors as a factor that makes the patron more likely to receive a promotion. If the operator answers “no,” the MGC requested information on what controls are in place that prevents that from happening.

Serpa’s letter requests each operator to detail what specific information is used for each individual patron to determine whether, when, or how that patron is marketed to. This included betting frequency, amount wagered, average wager size, types of bets placed, net winnings or losses, VIP or customer-value states, and several other patron factors that would influence the type or frequency of marketing.

The letter also asks each operator to provide information on how much these artificial intelligence and machine learning systems evaluate whether marketing or promotional models may encourage increased betting or contribute to problematic or addictive gambling, and also asks how a customer’s betting frequency, losses, and deposit activity causes the system to reduce, modify, or prohibit marketing to that patron.

Additional questions included in the letter:

• Does the Operator use increased wagering frequency or increased spending as a factor in determining whether a patron should receive additional or different marketing?
• How does the Operator determine which patrons receive a particular promotion or marketing communication?
• Does the Operator use machine learning to prediction a patrons likelihood of: Placing a wager? Increasing wagering activity? Returning to the platform? Depositing additional funds? Responding to a particular promotion? Increasing their spending?
• Can the system trigger a promotion or marketing communication based on a patron’s recent betting activity, losses, deposits, or other account activity?
• Does the Operator consider a patron’s gambling-risk indicators before allowing that patron to receive targeted promotions?
• Are there circumstances in which a patron’s betting frequency, losses, deposits, or other behavioral indicators cause the system to reduce, modify, or prohibit marketing to that patron? Conversely, are there circumstances in which those same indicators cause the system to increase or modify marketing directed to the patron?

In-Person Interviews Up Next

The MGC also plans to interview representatives from each operator in person on how artificial intelligence is used in such practices as customer acquisition, promotions, and responsible gaming. These interviews were scheduled to begin sometime this week.

Maynard said the commission will evaluate the interviews with its operators, provided written responses to the letter, and then potentially decide on future regulations regarding the use of artificial intelligence and machine learning by sports betting operators.

“Any further action or policy making can then be determined by the commission when and if appropriate,” Maynard said.