The Massachusetts Gaming Commission has instructed its staff to examine how sports betting operators deploy artificial intelligence and machine learning, a directive issued in the wake of a New York Times investigation into DraftKings. The report alleged the company used the technology in ways that shaped how customers were treated, prompting regulators in one of the more closely watched U.S. betting markets to take a harder look. The move places a spotlight on practices that, until now, most operators have kept behind closed doors.
What the Commission does with the findings could set a reference point for other state regulators weighing the same questions.
What the Review Means for Operators and Bettors
The regulatory response matters because Massachusetts is not a fringe jurisdiction. It is a licensed, mature market where compliance expectations carry weight, and where a formal staff evaluation can translate into concrete rulemaking. For the businesses involved, the implications are practical rather than abstract.
- Increased scrutiny of algorithmic decision-making: Operators may be required to explain how AI models influence customer limits, promotions, and account restrictions.
- Potential disclosure obligations: If the Commission moves toward rulemaking, licensees could face new transparency requirements around how machine learning tools are applied to individual accounts.
- Reputational exposure for any operator seen to be profiling customers rather than protecting them.
- A template for other states: Regulators elsewhere have signalled interest in AI oversight, and a Massachusetts framework gives them something to borrow from.
A Report, a Regulator, and a Technology Few Understand
The trigger was journalism, not a routine audit. The New York Times examined how the sports betting platform operated by DraftKings uses AI and machine learning, raising questions about whether the technology was deployed to influence customer behaviour in ways that favoured the operator. The Commission responded by directing staff to assess how these systems function across the licensed market, not just at a single company.
That distinction is worth holding onto. The review, at least on paper, targets an industry-wide practice rather than one firm. Machine learning has become standard infrastructure for sportsbooks: it prices markets, flags sharp bettors, tailors promotions, and manages risk in real time. But the same tools that optimise a book’s margin can also determine which customers get limited, which get bonuses, and which quietly get squeezed out.
And that is where the regulatory tension sits.
Why AI Oversight Is Harder Than It Looks
Regulating algorithms is not like regulating a payout percentage. A slot machine’s return-to-player figure is fixed, disclosed, and auditable. A machine learning model, by contrast, adapts continuously, drawing on thousands of data points to make decisions that even its operators may struggle to fully explain. Auditing such a system means auditing a moving target.
There is also the question of intent versus outcome. If a model restricts a winning bettor, is that a business decision, a fairness problem, or both? Regulators have been slow to draw that line, partly because the technology outpaced the rulebooks written to govern gaming. In practice, the Commission’s staff will need to decide what “acceptable” AI use even means before it can judge whether any operator crossed it.
Market Context and Regulatory Stakes
The following outlines the areas most likely to draw regulatory attention as the review proceeds.
| Area of Focus | Regulatory Concern |
|---|---|
| Customer profiling | Whether AI is used to identify and restrict profitable bettors while retaining losing ones |
| Promotional targeting | Whether personalised offers exploit behavioural patterns, including among at-risk or problem gamblers who may be algorithmically identified and then encouraged rather than protected |
| Transparency | Whether customers understand how automated decisions affect their accounts |
| Responsible gambling | Whether the same data used to boost engagement could instead flag harm |
The last row cuts to the core of it. The data that powers aggressive marketing is the same data that could power player protection. Which raises a harder question: if operators can identify a customer’s risk profile with precision, what obligation do they have to act on it?
How Operators May Respond
Expect measured, careful language from the industry. Sportsbooks are unlikely to abandon machine learning; it is too embedded in pricing and risk. Instead, the more probable path is voluntary disclosure of governance practices, internal model reviews, and public commitments to responsible use, moves designed to shape the regulatory outcome before it hardens into rules.
Not everyone will move at the same pace. Larger operators with mature compliance functions can absorb new requirements more easily than smaller competitors, which could tilt the market further toward scale. That is an outcome regulators rarely intend and rarely prevent.
Frequently Asked Questions
What did the Massachusetts Gaming Commission actually order?
It directed its staff to evaluate the use of artificial intelligence and machine learning by sports betting operators licensed in the state. The directive followed a New York Times report on DraftKings’ use of the technology.
Is DraftKings the only company under review?
No. While the report focused on DraftKings, the Commission’s evaluation is aimed at the practices of sports betting operators across the Massachusetts market.
Could this lead to new rules?
Possibly. A staff evaluation is a first step, not a regulation. But it often precedes formal rulemaking, particularly when the underlying concern touches customer fairness and responsible gambling.
Why does AI use in betting matter to regulators?
Because the same models that price bets and manage risk can also decide how individual customers are treated, from bonuses to restrictions. Regulators want to understand whether that decision-making is fair and transparent.
Does this affect bettors right now?
Not immediately. The review is at an early stage, and no changes to operator practices have been mandated as a result.
What Comes Next
The Commission’s staff will conduct its evaluation and report back, at which point the regulator can decide whether existing oversight is sufficient or whether new requirements are warranted. Other states are watching. The precedent set in Massachusetts gives them a starting framework, and a reason to ask their own operators the same uncomfortable questions.
