EFF Report: DraftKings Uses AI to Behaviorally Target Chronic Gamblers

In an article published on September 24, 2026, the Electronic Frontier Foundation (EFF) alleged that online sports betting company DraftKings is using AI to target customers who are most likely to lose bets and to respond to gambling promotions. EFF classifies this targeting as "online behavioral advertising," meaning companies use the user data they collect to personalize the ads they display.

According to what EFF relays from The New York Times, DraftKings uses customer gambling records to train a machine learning model to identify losing gamblers; once identified, targeted ads are sent to these customers to lure them back to the site to place more bets—bets that DraftKings believes will most likely lose. EFF further points out that DraftKings has a business incentive to keep losing gamblers coming back, because these users are the ones who actually generate revenue for it.

EFF also infers that people classified as "problem gamblers"—that is, people who gamble repeatedly despite harm already done to themselves, their finances, and their relationships—are highly likely to be targets of the model. In EFF's account, by reactivating these people through targeted promotions designed to keep them on the platform, DraftKings is profiting from their vulnerability rather than reducing their risk.

EFF acknowledges that predatory online behavioral advertising is not new, but argues that companies using AI to process data and target customers amplifies its harm. Its chain of logic is: online behavioral advertising itself incentivizes companies to collect massive amounts of data to support ad technology; once AI is introduced, even more data must be collected to train and optimize models.

A key reason EFF gives is the "black box" nature of AI: the people building a model can hardly predict which data points will be most useful to the AI, which drives companies to keep collecting more data. At the same time, AI lets companies process enormous datasets at far greater speed, thereby "amplifying" the harm of online behavioral advertising.

It should be noted that the above causal judgment is EFF's argument and inference; the source provides no quantitative evidence to show the extent to which AI amplifies the harm.

EFF places the DraftKings case within a larger data chain. It says the data collected for targeted advertising is being sold to insurance companies, banks, and state and federal law enforcement agencies including CBP (U.S. Customs and Border Protection).

EFF also says ICE (U.S. Immigration and Customs Enforcement) is paying attention to the data underpinning ad technology: earlier this year, ICE issued a request for information seeking to understand how commercial big data and ad tech providers can directly support investigative activities.

However, the source does not state the specific publication date or document number of ICE's request for information, nor does it state the specific scope and scale of advertising data being sold to insurance companies, banks, and law enforcement agencies. The actual degree of connection between ICE's request and advertising technology data is likewise unclear.

EFF says DraftKings appears to use only "first-party data" to target ads—that is, only data it collects directly from users, without purchasing additional data from third parties to train its machine learning model. EFF's use of the word "appears" indicates that this judgment itself carries uncertainty.

From this EFF draws a policy conclusion: regulatory proposals that only restrict third-party data sharing and sales are insufficient to stop this kind of predatory advertising. The implication is that even if third-party data sources are cut off, companies can still rely on the first-party data they hold to carry out the same targeting.

This inference is key to understanding EFF's argument: it shifts the debate from "where the data comes from" to "whether behavioral advertising itself should be allowed."

Based on the above argument, EFF advocates a full ban on online behavioral advertising. It argues that merely restricting third-party data sharing and sales is insufficient to stop such predatory advertising, and that policymakers must ban online behavioral advertising.

EFF's logic is: if companies cannot send personalized ads, their incentive to collect the behavioral data supporting those ads will diminish. EFF says DraftKings using AI to target losing gamblers is just one example of online behavioral advertising causing real harm to real people.

The source provides no empirical support to prove whether a full ban on online behavioral advertising would effectively stop such conduct; this claim currently remains at the level of EFF's policy position.

All substantive allegations in this article about DraftKings' conduct come from secondhand information relayed by EFF from The New York Times. No response from DraftKings appears in the source, nor is there any first-party evidence. EFF explicitly states that its claim that DraftKings used customer betting records to train a machine learning model to find losing gamblers comes from The New York Times.

The source provides no specific figures, such as the number of targeted users, model accuracy, revenue impact, or time span; it also does not explain the machine learning model's technical details, training data scale, or deployment time. In addition, the source does not provide the title, date, or link of the original New York Times article, making it impossible to verify its original reporting.

Therefore, whether DraftKings actually uses AI to target losing gamblers currently rests only on EFF's relaying of The New York Times, lacking independent verification; the extent to which "problem gamblers" are targeted by the model is EFF's inference, and the source gives no quantitative evidence; whether DraftKings uses only first-party data is described by EFF as "appears," which likewise carries uncertainty.