A New York Times investigation published September 19, 2026 revealed that DraftKings built a machine-learning model in 2023 that scored customers by how much money they were expected to lose after receiving a promotion. Customers with high “elasticity” scores, often the heaviest losers, were flooded with bonuses and free bets, while a separate AI project designed to detect problem gambling was deprioritized. The report drew on internal memos, betting records, and interviews with more than 40 current and former employees, and it has intensified calls for regulation of AI-driven gambling marketing. The findings highlight how DraftKings Used AI to Target Biggest Losers.
What Is DraftKings’ AI Promotion Targeting System?
DraftKings’ AI promotion targeting system is a machine-learning model, built in 2023, that predicted which customers would bet and lose the most money after receiving a bonus or free bet. The core output was an internal “elasticity” score: the higher the score, the more money DraftKings projected that gambler would lose after receiving an offer, as reported by the Times.
This investigation illustrates the controversial tactics that DraftKings Used AI to Target Biggest Losers, raising ethical concerns in the gambling industry.
The model worked by feeding detailed customer betting records into an algorithm. Former data analyst Jayden Butts tested the model beginning in 2023, first on roughly 5,000 online casino users around September 2023 before a broader rollout, according to follow-up coverage. The trials reportedly validated that targeting high-elasticity users increased both betting volume and net losses.
How Does DraftKings Use AI to Identify Problem Gamblers?
Here’s the sharpest finding in the reporting: DraftKings did build AI to identify problem gamblers, and then reportedly let it stall. Data scientist Nestor Hernandez developed a separate machine-learning model intended to flag signs of gambling problems or addiction risk in customer behavior, according to the Times’ reporting.
Former employees say that harm-detection project was later scaled back, delayed, or effectively shut down, even as the promotional elasticity model was refined and expanded. That contrast is the heart of the controversy: one AI system, designed to maximize revenue from heavy losers, was fully deployed. Another, designed to catch harm early, was not.
DraftKings disputes the framing. The company says it operates a responsible-gambling system that monitors more than two dozen indicators of risky behavior, rapid losses, frequent deposits, chasing losses, and that promotions aim at customers who show “sustained, engaged use” of its products, as Yahoo’s coverage summarizes.
Is It Legal for Sportsbooks to Target Losing Customers With Promotions?
In most U.S. states, no law currently prohibits a sportsbook from using AI to direct promotions at customers likely to lose money. That’s the legal gap this investigation exposes. Regulated operators must offer responsible-gambling tools, but the law generally doesn’t restrict how they target their marketing.
That may be changing. Policy observers note that steering offers toward the “biggest losers” could prompt investigations into unfair or exploitative practices in AI-driven gambling marketing, and industry commentary suggests regulators in multiple jurisdictions are already examining how sportsbooks use data and AI.
The DraftKings “Biggest Losers” Promotion Controversy Explained
The controversy, in plain terms: DraftKings reportedly used the same behavioral signals that indicate gambling harm, frequency, escalating losses, balance patterns, to sell more gambling rather than to intervene. The model incorporated gambling frequency, daily account balances, loss-to-wager ratios, and dozens of other metrics derived from casino and sports-betting histories, according to TechTimes’ analysis of the investigation.
Customers labeled “inelastic”, less responsive to offers, received fewer promotions. High-scoring, heavy-loss bettors were repeatedly targeted. As of September 2026, the elasticity model reportedly remains part of DraftKings’ marketing machinery, while the dedicated problem-gambling detection model has not been similarly deployed, as AI-focused coverage notes.
How Much Money Do Problem Gamblers Lose, and How Big Is the Promotional Push?
The reported scale is substantial. Coverage of the Times investigation indicates roughly $400 million in AI-driven promotions in 2025 was directed at users the model identified as most lucrative. That figure comes from industry reporting on the internal promotional budget.
On individual losses, the reporting centers on betting records and internal projections rather than a single aggregate figure. What former employees describe is a system optimized so that promotional spending flowed to the customers whose expected losses were highest, meaning the heaviest losers effectively subsidized their own escalation with “free” bets.
What Other Sportsbooks Use Similar AI Targeting Tactics?
DraftKings is not an outlier. Industry and tech-press analyses stress that major online gambling operators increasingly use AI and predictive analytics to identify “high-value” customers, often those who lose consistently, and concentrate marketing and bonuses on them, as Moneycontrol’s report explains.
The difference is the paper trail. The Times obtained DraftKings’ internal memos and presentations, making this the best-documented public case of a U.S. sportsbook’s algorithmic targeting. Commentators warn the practice is likely industry-wide; DraftKings is simply the one caught on record.
What Are the Risks of Personalized Betting Promotions?
Personalized promotions can turn a person’s own data into a weapon against them. Experts and consumer advocates cited in the coverage argue that using AI to maximize losses from vulnerable bettors is fundamentally at odds with responsible-gambling promises.
The specific risks:
- Exploiting loss-chasing. Signals like chasing losses or increasing bet sizes, classic harm markers, become marketing triggers rather than intervention flags.
- Masking harm. A “free bet” looks like a perk, but for a high-elasticity customer it’s engineered to deepen losses.
- Asymmetric information. The company knows your elasticity score. You don’t.
- Undermining safeguards. If harm-detection tools are deprioritized, self-exclusion and limits arrive too late.
Can You Opt Out of Targeted Gambling Promotions?
Yes, though it takes deliberate effort. You can reduce or stop promotional contact through several channels:
- Notification settings. In the DraftKings app and website account settings, you can opt out of promotional emails, push notifications, and SMS offers.
- State opt-out lists. Some states maintain exclusion or self-exclusion programs that bar marketing contact from licensed operators.
- Direct requests. Federal rules generally require companies to honor requests to stop marketing messages.
- Carrier-level blocking. Filtering tools can block gambling-related marketing texts and emails.
The important caveat: opting out of promotions doesn’t disable the underlying scoring. Self-exclusion is the stronger step if gambling has become a problem, because it closes the account entirely for a set period.
What Responsible Gambling Tools Does DraftKings Offer?
DraftKings offers deposit limits, wagering limits, time limits, time-outs, and self-exclusion, plus a responsible-gambling system the company says monitors more than two dozen risk indicators. Critics’ point, drawn from the Times investigation, is not that these tools are absent, it’s that the company reportedly invested more in the AI that identifies profitable losers than in the AI that identifies at-risk gamblers.
What Should Regulators Do About Sportsbook AI Targeting?
The reasonable middle ground most advocates propose: don’t ban AI, but ban its use against the vulnerable. Concrete steps under discussion include:
- Restrict targeting based on harm indicators. Signals like chasing losses or escalating deposits should trigger intervention, not bonuses.
- Require parity. If a company builds a revenue-optimization model, it should maintain an equally resourced harm-detection model.
- Mandatory algorithmic audits. Independent reviews of how promotional AI scores customers.
- Marketing freezes for flagged accounts. A customer showing risk markers should receive no promotions at all.
- Public disclosure. Bettors deserve to know when an offer is algorithmically targeted.
For New York readers, this matters directly. Sports betting is legal and heavily marketed across the state, including here in the Mohawk Valley, and the New York State Gaming Commission is among the regulators that could act on AI-targeting practices. Contacting state representatives to support algorithmic-audit requirements is one practical step.
FAQ
What did the Times find about DraftKings?
A September 19, 2026 New York Times investigation found DraftKings used a machine-learning model to steer promotions toward customers most likely to lose money, based on internal documents and interviews with more than 40 current and former employees.
What is DraftKings’ elasticity score?
It’s an internal metric that ranked each customer by expected loss after receiving a promotion. Higher scores meant more targeted bonuses; low-scoring customers received fewer offers.
Did DraftKings break the law?
No regulator has accused DraftKings of illegality as of this reporting. The controversy centers on whether the practice is exploitative and whether existing law adequately covers AI-driven targeting.
How much did DraftKings spend on AI-driven promotions?
Coverage of the investigation reports roughly $400 million in AI-driven promotions in 2025 aimed at users the model identified as most lucrative.
What has DraftKings said?
The company says promotions target customers with “sustained, engaged use,” not people with gambling problems, and cites a responsible-gambling system monitoring more than two dozen risk indicators.
Can I stop receiving DraftKings promotions?
Yes. Opt out through app notification settings, contact the company directly, or use state self-exclusion programs, which bar marketing from licensed operators.
Key Takeaways
- DraftKings developed an internal “elasticity” score ranking customers by expected losses after promotions, according to the Times investigation.
- Roughly $400 million in AI-driven promotions in 2025 were steered toward users the model flagged as most lucrative.
- A parallel machine-learning project meant to detect signs of gambling addiction was reportedly scaled back while the promotions model expanded.
- DraftKings says its promotions target customers with “sustained, engaged use,” not people with gambling problems, and points to a responsible-gambling system monitoring more than two dozen risk indicators.
- Experts say using addiction-adjacent behavior markers for marketing is fundamentally at odds with responsible-gambling commitments.
- Regulators in multiple jurisdictions are already examining how sportsbooks use data and AI.
Conclusion
The DraftKings story is ultimately about priorities. The company had the talent to build AI that detects problem gambling, and reportedly chose to perfect the AI that profits from it instead. That’s not a technology problem. It’s an accountability problem, and it’s exactly the kind of corporate behavior that demands government transparency and regulatory guardrails.
Here’s what readers can do: check your own betting app’s promotional settings today, use deposit and time limits if you gamble at all, and contact your state representatives to support algorithmic-audit and harm-parity requirements for sportsbooks. Informed customers and engaged citizens are the strongest check on practices that no regulator has yet caught up to.
