DraftKings Accused of Using AI to Target Losing Gamblers, Shelved Internal Safety Model
The algorithm finds the biggest losers, then sends them coupons to keep betting.
Online sports betting giant DraftKings is facing regulatory scrutiny and a federal lawsuit filed on October 1 in Boston, following investigations into its algorithmic marketing practices. Since 2023, the company has deployed a machine learning model that analyzes customers' betting frequency, daily account balances, and net losses to calculate an 'elasticity score.' The algorithm is designed to identify users who are most likely to respond to promotional offers by placing additional losing bets, automatically targeting them with personalized incentives to keep them on the platform.
An investigation by ProPublica journalist Jake Pearson demonstrated the system in practice: after losing nearly $1,800 on basketball bets in a single evening, rather than receiving a responsible gaming intervention, DraftKings immediately invited him to try out its VIP program. According to former employee Jayden Butts and more than 40 former staff members interviewed by The New York Times, internal data scientists had also built a counterpart AI model capable of detecting compulsive gamblers in crisis and offering help. However, DraftKings leadership canceled a scheduled demonstration and killed the safety project entirely.
According to the World Health Organization, individuals with harmful gambling behaviors generate roughly 60% of all gambling losses. The Electronic Frontier Foundation (EFF) warned that AI is 'supercharging the harms of online behavioral advertising' by weaponizing personal behavioral data against vulnerable individuals. The Massachusetts Gaming Commission has opened an inquiry into DraftKings' AI practices, alongside new class-action litigation accusing the company of weaponizing machine learning to profit off addiction.
Sources: The New York Times, ProPublica, Electronic Frontier Foundation, The Boston Globe, Malwarebytes.
BenPig is holding its piggy bank a little tighter.