DraftKings Is Using AI to Supercharge the Harms of Online Behavioral Advertising

There's a particular flavor of dystopia that arrives not with a bang, but with a terms-of-service update. The EFF's report on DraftKings reveals something far more insidious than simple targeted ads: an AI system engineered to identify the precise psychological and behavioral signatures of a losing bettor. In the speculative fiction I write about, we imagine AIs that can predict human weakness—Skynet's psychological warfare, the precogs of *Minority Report*. We rarely stop to consider that the technology is already here, just wearing a sportsbook's logo and a corporate privacy policy. What makes this development uniquely chilling is the inversion of the advertising paradigm. Traditional advertising seeks out those most likely to buy; DraftKings' AI seeks out those most likely to *lose*. The system isn't finding customers—it's finding victims. It analyzes behavioral data to determine which users are on tilt, which ones are chasing losses, which ones have the addictive personality profile that responds to "free bet" promotions despite a track record of failure. This is the difference between a street hustler running a shell game and a data scientist running a neural network designed to separate you from your money with surgical precision. The broader theme here is the quiet normalization of algorithmic predation. We've accepted behavioral advertising as the price of a "free" internet, but EFF has long warned that the same infrastructure can be pointed at increasingly harmful ends. The URL—https://www.eff.org/deeplinks/2026/09/draftkings-using-ai-supercharge-harms-online-behavioral-advertising—documents not just DraftKings' practices, but the systemic failure of a regulatory framework that treats gambling ads like cereal commercials. When an AI can calculate the exact moment you're most likely to make a desperate bet, the line between persuasion and exploitation has been not merely crossed, but obliterated. Looking forward, this is a harbinger. If an AI can be trained to find the losing bettor, it can be trained to find the financially desperate, the medically anxious, the emotionally fragile. The same architecture that DraftKings uses to target gamblers could be repurposed by payday lenders, by "wellness" scams, by political operatives. The EFF's report is a warning shot, but the question is whether we're listening—or just placing another bet that the future will be kinder than the present.