DraftKings is facing a proposed nationwide class action that claims the company used artificial intelligence to identify the bettors most likely to lose money, then buried them in promotions. Daniel Vest, a West Virginia resident, filed the suit September 30 in the U.S. District Court for the District of Massachusetts. At the center of his claim is a number: roughly 70 promotional emails, texts and push notifications in the 30 days before September 26.
The complaint targets an internal machine-learning model built in 2023 that scored online casino players on an "elasticity" metric, essentially a prediction of how much more each customer would wager and lose after receiving a free bet or bonus. High scorers got more offers. With about $3 billion flowing into promotions against $8.7 billion in gross gaming revenue, the model had plenty of budget to work with.
What the suit claims, and what DraftKings says
Vest's legal team frames it bluntly, arguing DraftKings "weaponized AI" against its own customers. The filing says the company breached privacy notices that promised user data would support responsible play, and it brings unfair and deceptive practice claims under Massachusetts law. Vest says he lost thousands of dollars over years of betting. The requested remedies include damages, disgorgement of promotion-driven revenue and an injunction against the targeting.
DraftKings rejects the premise. The company says promotions go to "customers who demonstrate sustained, engaged use of our platform, not toward customers based on their losses," and it dismisses one internal test highlighted in the filing as preliminary and inconclusive. It also states plainly that it doesn't use AI to target players based on losses or problem-gambling indicators.
Regulators aren't waiting for a verdict. The Massachusetts Gaming Commission opened an investigation into AI-driven promotional targeting in late September, with chair Jordan Maynard directing staff to engage the company directly. DraftKings is headquartered in Boston, so the home-state pressure carries real weight.
The detail that should worry every bettor
Here's the part that got less attention. Internal documents and interviews with more than 40 former employees describe a harm-detection model that was actually built, then sidelined because the company didn't consider it "evidence-based." The revenue model shipped. The safety model sat on a shelf. Former data analyst Jayden Butts summed up the internal logic in a line that's hard to unread: "Is this person going to give us more than we're giving them? Open the floodgates."
I spent time in the Reddit and Hacker News threads once the internal documents surfaced, and the reaction isn't the usual sportsbook grumbling. Users call the practice predatory and bleak, and the sharpest comments target the gap between responsible gaming branding and algorithmic loss-chasing in the actual product. There's also a recurring personal note, people asking why their promo volume spiked exactly when their betting got heavier.
That tracks with what readers describe in our DraftKings review, where promo cadence is among the most common complaints. Others are pushing for the SAFE Bet Act or state bans on algorithmic targeting.
The privacy angle deserves more attention than it's getting. Nothing here involved data brokers or third-party sharing. Every signal the model consumed, bet frequency, balances, loss ratios, came from DraftKings' own first-party records. Most privacy rules target data that moves between companies. Data that never leaves the building sits largely outside that framework, which is why consumer groups argue behavioral targeting itself needs new rules.
And notice the asymmetry in how the tools work. Deposit limits and self-exclusion are opt-in. The elasticity scoring ran automatically, and users had no way to see their score or decline to be scored. You can volunteer to be protected, but you can't volunteer to not be targeted. Whether that scoring still operates the same way today isn't public knowledge, and that gap matters.
What happens next
The lawsuit faces real hurdles. It's a nationwide class built on one plaintiff's experience, filed under Massachusetts consumer law, which raises choice-of-law questions for bettors in other states. Proving that individual losses trace to algorithmic targeting rather than to gambling itself will be the central fight, and DraftKings' denial hands the company a clean defense for now.
Discovery is the thing to watch. If the case survives early motions, internal documents on elasticity thresholds and the shelved harm model could become public in a way no investigation has forced yet. Every operator running a bonus engine, from DraftKings to European books like Betway, has reason to follow each filing closely. We track promo mechanics across markets for exactly this reason.
My own read is that the courtroom outcome matters less than the asymmetry already on the record. A company built a model to predict who would lose the most, and built a model to flag who was being harmed, then funded one and shelved the other. Whatever a judge decides about Massachusetts consumer law, that's the sentence every state regulator should have to respond to.