DraftKings AI Lawsuit: Bettor Sues Over 'Elasticity' Model That Allegedly Targeted Likely Losers

The DraftKings AI lawsuit is a proposed nationwide class action filed on September 30, 2026 in the United States District Court for the District of Massachusetts by Daniel Vest, a West Virginia bettor. It alleges that DraftKings used a machine-learning "elasticity" model to find customers likely to keep gambling and losing, then flooded them with promotions without telling them. The case matters because it does not depend on proving addiction or a duty of care. It frames AI-targeted marketing as a breach of DraftKings' own privacy promises and, potentially, as an unfair practice under Massachusetts Chapter 93A, all against the backdrop of a state rule that already restricts addictive AI-driven promotions.

DraftKings denies the core allegation. "DraftKings does not use AI to target customers based on losses, nor do we use AI to target customers based on indicators of potential problem gaming," the company said in a statement reported by WBUR, Casino.org and others, adding that it intends to defend vigorously. No court has ruled on anything yet. The questions now are whether the claims survive a motion to dismiss and whether DraftKings' arbitration clause moves the case out of court. If it does, the next question is whether thousands of individual demands follow.

What is the DraftKings AI lawsuit about?

The DraftKings AI lawsuit alleges that the Boston-based sportsbook and online casino built a model that predicted how much more each customer would wager and lose in response to a promotion, and then aimed its heaviest marketing at the customers the model rated most responsive. The complaint borrows the company's alleged internal vocabulary: customers rated above average were "elastic" and received more emails, texts, push notifications and casino bonus offers, and customers rated below average were "inelastic" and slated for fewer offers.

Vest says he bet thousands of dollars a year on DraftKings for several years. He alleges he received about 70 emails, texts, app notifications and other messages in the roughly 30 days ending September 25, 2026. The Boston Globe quotes the complaint's central charge: DraftKings "has weaponized AI to do just that — understand and exploit users' vulnerabilities to financially benefit the company." The theory of harm is nondisclosure plus misuse of data. Vest says he was never told that his deposits, balances and losses were feeding a model built to maximize what he would spend.

The filing came 11 days after a New York Times investigation that the complaint reportedly quotes throughout. That reliance matters procedurally. At the pleading stage, Vest needs plausible factual allegations, not proof, and a detailed published investigation based on internal documents gives him more specific allegations than a typical consumer complaint has.

Who filed the DraftKings AI lawsuit and where?

Daniel Vest, a West Virginia resident and long-time DraftKings customer, filed the case as the Named Plaintiff in federal court in Boston, where DraftKings has its headquarters. The only defendant identified in published reports is DraftKings Inc. Filing in DraftKings' home district avoids personal jurisdiction fights for a nationwide class and puts the case before judges who have seen DraftKings' terms of use before (discussed below).

Federal jurisdiction will almost certainly rest on the Class Action Fairness Act (CAFA): a proposed class of more than 100 members, minimal diversity (a West Virginia plaintiff against a Massachusetts-based company), and an aggregate amount in controversy above $5 million, which a nationwide class of bettors will readily allege. Published reports have not identified the docket number or plaintiff's counsel.

What is DraftKings' elasticity score?

DraftKings' elasticity score, as described in the New York Times investigation published September 19, 2026, is a machine-learning estimate of how much additional revenue a customer generates for each promotional dollar the company spends on them. Coverage of the Times report says the score drew on how often a customer plays, how the account balance moves, how much the customer bets and loses, and how likely the customer is to stop gambling. According to that coverage, the casino model was built in 2023 to replace a cruder rules-based system, and in 2025 DraftKings used it to direct roughly $400 million in bonus bets and perks toward its highest scorers.

The Times based its reporting on interviews with more than 40 former employees, internal memos and presentations, Slack messages and customer betting records. It also reported that a separate machine-learning project meant to flag signs of problem gambling was scaled back or shelved while the promotional model was refined. That contrast between a harm-detection tool that was set aside and a revenue tool that was expanded is likely to be the plaintiff's strongest theme. It is also the part of the story most exposed to discovery: if the internal documents exist in the form described, they are discoverable in the Massachusetts case.

DraftKings told the Times it rejects any suggestion that its marketing is unfair or improperly targets customers. Its public statement on the lawsuit is narrow. It denies targeting customers "based on losses" or "based on indicators of potential problem gaming." It does not deny using machine learning to allocate promotions. Expect that distinction to shape the motion to dismiss: an operator can argue that modeling responsiveness to offers is ordinary marketing analytics, not loss-targeting.

What claims does the DraftKings AI class action make?

The DraftKings AI class action pleads three common-law counts: breach of contract, breach of implied contract, and money had and received. This is according to Legal Sports Betting's review of the complaint, and other outlets describe the claims more generally as an unfair and deceptive practice and a breach of DraftKings' contractual obligations. The contract counts rest on the "Responsible Play" passage in DraftKings' sportsbook and casino privacy notices, in which the company says it may review customer information for signs of irresponsible play and reach out with resources. Vest's theory is that DraftKings used the same data "for the exact opposite purpose," to identify customers who would lose more.

The Chapter 93A piece is the one to watch. A footnote in the complaint reportedly says Vest is not pressing a Chapter 93A claim "at this time." It says a demand letter under Section 9 was sent with the filing and gives DraftKings 30 days to make "an acceptable tender of settlement." If DraftKings does not, Vest plans to move to add the 93A claim and seek additional damages. This sequencing follows the statute. Chapter 93A, Section 9 requires a written demand at least 30 days before suit. A defendant that rejects the demand with a tender the court later finds unreasonable faces a remedy enhancement. For a willful or knowing violation, or a bad-faith refusal to make a reasonable tender, the court awards double or treble damages plus attorney's fees. Section 9 also expressly allows class actions.

The 93A route also connects the lawsuit to the Massachusetts gaming rules. Under the Attorney General's regulation at 940 CMR 3.16, failing to comply with a statute or regulation meant to protect the public's health, safety or welfare can itself be an unfair or deceptive act. A plaintiff who can plausibly allege a breach of 205 CMR 257.02 therefore has a path to a 93A claim even though the gaming regulation does not itself create a private right of action.

Who is in the proposed class in the DraftKings AI lawsuit?

The proposed class in the DraftKings AI lawsuit covers every DraftKings customer, in any state, whom the company deemed "elastic" or sent targeted messages or promotions based on the model's output within the statute of limitations. The relief sought includes damages, repayment or disgorgement of money DraftKings obtained through the practice, and an injunction barring the company from using the model to push more gambling promotions.

The definition anticipates two predictable Rule 23 fights:

• Ascertainability. Because membership is tied to DraftKings' own scores and campaign logs, the plaintiff can argue the class list already exists in the defendant's data. That is an unusual advantage for a consumer class.
• Predominance. DraftKings will argue that individual questions overwhelm common ones: which privacy notice each customer saw, whether a promotion caused any particular bet, and whether a customer won or lost on balance. A contract theory built on a uniform form notice is the plaintiff's answer, because breach can be shown with common proof even if damages vary.

The nationwide scope raises choice-of-law questions too. DraftKings' terms and notices likely include a governing-law clause, and Massachusetts courts generally enforce those. If Massachusetts law governs the contract counts for the whole class, the multistate problem shrinks. If not, Class Certification becomes harder.

Does Massachusetts law ban AI-targeted gambling promotions?

Massachusetts does not ban AI in sports betting marketing, but 205 CMR 257.02 bars a sports wagering operator from using patron data to promote wagers or offers through "any computerized algorithm, automated decision-making, machine learning, artificial intelligence, or similar system that is known or reasonably expected by the Sports Wagering Operator or a vendor to the Sports Wagering Operator to make the gaming platform more addictive." The Massachusetts Gaming Commission approved the data privacy rules in 2023, and Commissioner Paul Brodeur has said Massachusetts may be the only state whose rules mention AI at all.

The same section limits use and retention of patrons' confidential and personally identifiable information to legitimate business purposes needed to operate or advertise a sports wagering platform, or to comply with law. For any use beyond those purposes, the operator must get the patron's consent. That consent must be clear, conspicuous and separate from other agreements, and patrons must be able to withdraw it at any time. A plaintiff will argue that loss-sensitive promotional modeling is a use that needed separate consent. DraftKings will argue that it is ordinary advertising of the platform.

The "known or reasonably expected" standard is the hinge. It does not prohibit personalization. It prohibits personalization the operator knows or should expect will make the platform more addictive. Internal documents about a shelved problem-gambling model could bear directly on what DraftKings knew or reasonably expected.

The regulator has not made findings. On September 24, 2026, MGC Chair Jordan Maynard asked Executive Director Dean Serpa and staff to engage with DraftKings on the specifics the Times reported and to examine how all licensed operators use these technologies. Any further action or rulemaking will be decided by the commission later, if at all. The commission has not determined that DraftKings violated any Massachusetts regulation.

Can DraftKings force the AI lawsuit into arbitration?

DraftKings can try, and history in the same courthouse favors it. DraftKings' terms of use contain a mandatory individual Arbitration Clause with a Class Action Waiver. A third-party terms tracker summarizes the current version as requiring JAMS arbitration and offering a 30-day Opt-Out Clause from account creation. In 2019, Senior Judge George A. O'Toole Jr., presiding over the daily fantasy sports Multidistrict Litigation (MDL) in the District of Massachusetts (MDL No. 2677), compelled players' class claims against DraftKings and FanDuel into arbitration. He reached even some plaintiffs without their own agreements because their claims were intertwined with arbitrable ones. He left challenges to the terms as a whole for the arbitrator.

Published coverage of Vest's complaint does not say how it deals with the arbitration clause. There are a few common ways a plaintiff tries to stay in court:

1. Opt-out. A customer who opted out within the window is not bound. That helps only the opted-out customers, and the class definition would have to be narrowed to them or split.
2. Formation. Massachusetts applies a two-part test for online terms from Kauders v. Uber Technologies (2021): reasonable notice of the terms and a reasonable manifestation of assent. Older sign-up flows and app updates are where formation challenges succeed.
3. Scope and claims about the privacy notice. The plaintiff may argue that promises in a separate privacy notice fall outside the arbitration clause, or that the injunctive relief sought reaches beyond any individual dispute. A Delegation Clause incorporating JAMS rules would likely send that scope fight to the arbitrator.

If DraftKings files a Motion to Compel Arbitration and wins, the class case would likely be stayed and the claims pushed into individual JAMS proceedings. That is where the economics change. A nationwide pool of "elastic" customers, if counsel can identify and sign them up, is a ready-made Mass Arbitration inventory. Each demand would trigger business-side Filing Fee and arbitrator costs under JAMS's consumer standards and its mass-arbitration procedures, and DraftKings would likely respond with Batching or bellwether proposals. FanDuel's 2026 win compelling a $250 million gambling-loss claim into individual arbitration shows how operators use these clauses. The Chapter 93A fee-shifting and multiple-damages provisions would follow the claims into arbitration if Massachusetts law governs.

How the DraftKings AI lawsuit differs from earlier gambling addiction suits

Most US sports betting addiction suits so far have pleaded negligence or product liability. They argue an operator owed a duty to protect a compulsive bettor, or that the app is a defective product. Those theories have struggled. In Macek v. DraftKings, a federal judge in the Eastern District of Pennsylvania held on March 23, 2026 that DraftKings has no duty of care to protect bettors from spending too much or from developing a gambling addiction. That decision is on appeal to the Third Circuit. Individual suits such as the September 16, 2026 VIP-manager complaint against DraftKings and FanDuel also face individual Arbitration Clauses.

Vest's complaint avoids the duty question. It does not need a court to recognize a new tort duty. It asks the court to enforce what DraftKings allegedly promised in its own privacy notices, and it sets up a statutory unfairness claim in a state whose gaming rules mention AI by name. It also does not require proof of clinical addiction or individual causation for liability. The class only needs to show that the company used customer data in a way the notices did not allow. That makes it a consumer protection and data-use case rather than a Mass Tort, though plaintiffs' firms building gambling-harm inventories will watch it closely.

Timeline of the DraftKings AI lawsuit

Date	Event
2014	Arbitration provision first added to DraftKings' terms of use, per a law review account
2019	Judge O'Toole compels DFS players' class claims against DraftKings and FanDuel into arbitration (D. Mass., MDL No. 2677)
2023	MGC adopts 205 CMR 257 sports wagering data privacy rules, including the AI clause; DraftKings builds the casino elasticity model, per the Times
2025	About $400 million in bonus bets and perks directed to the highest scorers, per the Times
March 23, 2026	Macek v. DraftKings (E.D. Pa.): no duty of care to prevent gambling losses or addiction
September 16, 2026	VIP-manager suit filed against DraftKings and FanDuel in E.D. Pa.
September 19, 2026	New York Times investigation of DraftKings' elasticity model published
September 24, 2026	MGC directs staff to review DraftKings' and other operators' AI use
September 30, 2026	Vest files the proposed class action in the District of Massachusetts; Chapter 93A demand served
About October 30, 2026	End of the 30-day window for a settlement tender before a motion to add the 93A claim

What it means for claimants' counsel, operators and funders

For plaintiffs' and claimants' counsel, the case offers a template that avoids the duty-of-care problem. It pleads breach of the operator's own data promises, uses state gaming rules through Chapter 93A, and defines the class by the defendant's own model outputs. The weak point is arbitration. Counsel should expect a motion to compel early and should check opt-out status and sign-up flows before relying on a class. Firms already holding sports betting intake should note that a contract and data-use theory may support claims from customers who cannot prove addiction.

For sportsbook operators and their counsel, the practical exposure is in the documents. Privacy notices that promise responsible-play monitoring create a contractual baseline that marketing models will be measured against. Massachusetts licensees should expect MGC staff to ask how promotional models are built, what they use as inputs, and whether any responsible-gaming model was considered and dropped. Operators in other states should review whether their notices promise more than their marketing systems deliver. Clause drafters should look again at Mass Arbitration protocols, because a class defined by a model score can turn into an arbitration inventory overnight.

For litigation funders, the case has an unusual profile. Liability turns on common documents and a uniform notice, but damages per customer may be modest and arbitration risk is high. Third-Party Litigation Funding for a mass-arbitration strategy would depend on the JAMS fee structure and on whether Chapter 93A multiple damages and fee-shifting apply.

What to watch next in the DraftKings AI lawsuit

• DraftKings' response to the 93A demand before the 30-day window closes around October 30, 2026, and any motion to amend to add the Chapter 93A claim.
• A motion to compel arbitration or to dismiss, which will show whether DraftKings relies on the arbitration clause, attacks the contract theory, or both.
• MGC staff findings from the review announced September 24, 2026, and any rulemaking on AI in promotions.
• Copycat filings in other states or by state attorneys general, and any attempt to centralize them before the Judicial Panel on Multidistrict Litigation (JPML).
• The Third Circuit's decision in Macek, which will affect the duty-of-care cases that run alongside this contract theory.

FAQ

What is the DraftKings AI lawsuit about?

It is a proposed class action, filed September 30, 2026 in federal court in Boston, alleging DraftKings used a machine-learning "elasticity" model to identify customers likely to keep betting and losing and then targeted them with promotions without disclosing it.

Who filed the DraftKings AI lawsuit?

Daniel Vest, a West Virginia DraftKings customer who says he received about 70 promotional messages in roughly 30 days, filed it in the United States District Court for the District of Massachusetts on behalf of a proposed nationwide class.

What did DraftKings say about the AI lawsuit?

DraftKings said it "does not use AI to target customers based on losses, nor do we use AI to target customers based on indicators of potential problem gaming," and that it intends to defend vigorously.

Who would be in the DraftKings AI class action?

The proposed class covers every DraftKings customer the company labeled "elastic" or sent targeted messages or promotions based on the model's output within the limitations period, in any state. No class has been certified.

Can DraftKings force the case into arbitration?

Possibly. DraftKings' terms contain an individual arbitration clause and class action waiver, and the District of Massachusetts compelled arbitration of earlier fantasy sports class claims against DraftKings in 2019. Opt-outs, contract formation and scope arguments are the main ways to stay in court.

Does Massachusetts prohibit AI-targeted betting promotions?

Massachusetts regulation 205 CMR 257.02 bars operators from using patron data with AI or machine learning "known or reasonably expected" to make the platform more addictive, and requires separate consent for data uses beyond operating and advertising the platform.