DraftKings Under Fire: AI Models Allegedly Target High-Loss Gamblers

The use of Artificial Intelligence (AI) in online gambling has reached a new and highly controversial stage. While operators publicly promote responsible gaming, internal reports from US industry giant DraftKings paint a different picture. Former employees report a targeted hunt for customers who, statistically, lose the most money. Jayden Butts, a former data analyst at the company, revealed details about a machine learning model programmed to identify so-called elastic players. These are customers who react particularly strongly to promotional offers by wagering even more money and ultimately losing it.

The ethical dilemma is obvious: the individuals who are most profitable for the company often show the clearest signs of emerging or advanced gambling addiction. Butts worked on tests where free bets and bonuses were prioritized for these losers. His concerns were apparently ignored internally. The logic behind it is as simple as it is terrifying: from a purely financial perspective, a player at high risk for problem gambling is the best investment for the operator because they have the least control over their losses.

Numbers and facts

DraftKings is a powerhouse in the US market with around 11 million customers in 2026, a massive increase from 5 million in 2022. The company generated approximately $8.7 billion in Gross Gaming Revenue (GGR) last year, while simultaneously spending about $3 billion on promotions. Internal memos prove that data science helped improve margins on promotion-driven sports bets by 13 percent in 2025. A particularly drastic example of aggressive marketing is the case of Bryan Biehl. While he was in therapy for his addiction, he received 40 promotional emails from DraftKings in the first two weeks of December 2024 alone.

"It is as predatory as it sounds. If you lose more, we give you more, so you keep playing more." - Anonymous former DraftKings analyst

Background

The technological arms race at DraftKings is not limited to sports betting. One-third of the revenue now comes from online casino games like slots, which are considered particularly addictive. An internal 2023 memo noted that slot revenue was more elastic than earnings from other games. This means customers are particularly susceptible to marketing incentives in slots and lose disproportionately more of their own money after receiving a bonus. While marketing AI flourished, a project by data scientist Nestor Hernandez, which envisioned a risk-scoring system for early addiction detection, was reportedly shut down or paused according to former colleagues. Lori Kalani, Chief Responsible Gaming Officer at DraftKings, defends the practice as a standard marketing tool comparable to that of Amazon.

Why it matters for German players

For German players, this report is a stark warning against operators without a German license. In Germany, the Interstate Treaty on Gambling 2021 (GlüStV 2021) regulates the market extremely strictly to prevent exactly such algorithmic excesses. Anyone playing with a provider on the official whitelist of the Gemeinsame Glücksspielbehörde der Länder (GGL) is protected by technical barriers. The cross-operator deposit limit of 1,000 euros per month and the LUGAS system prevent players from falling into an endless downward spiral fueled by AI-driven bonuses. Furthermore, a stake limit of 1 euro per spin applies to virtual slots in Germany, which massively restricts the elasticity and exploitation of loss potential.

What it means for GGL-licensed casinos

Casinos with a GGL license must submit strict social concepts. A procedure like the one described at DraftKings targeted targeting of losers to increase revenue would lead to the immediate revocation of the permit in Germany. German authorities demand that algorithms be used primarily for player protection. Early detection systems for problematic gambling behavior are mandatory in Germany, not optional. The GGL also monitors advertising guidelines very closely; aggressive bonus floods, like those experienced by Bryan Biehl in the USA, are legally impossible under strict German advertising restrictions. German customers should therefore strictly look for the GGL seal to avoid becoming the target of such predatory algorithms.