Former DraftKings employees detail how it uses ML to target likely losers with promotions, while efforts to flag problem gamblers were shelved or squashed
About a year into his job as a data analyst at DraftKings, Jayden Butts received a new assignment.
Context & Ripple Effects
DraftKings faced integrity and regulatory scrutiny in 2015, including a New York attorney general inquiry into its fraud safeguards and employee-play restrictions adopted with FanDuel. The new allegations shift the focus from protecting contest integrity to how customer data is used to allocate promotions.
Former employees’ accounts describe a tension within DraftKings’ data operations: models were used to identify customers likely to lose, while problem-gambling detection efforts were reportedly shelved or suppressed. That makes promotional targeting, rather than just the existence of machine learning, the central business and consumer-protection issue.
First-order effects
- DraftKings customers identified as likely losers are more likely to receive betting incentives, according to former employees, concentrating promotional spend on a defined behavioral segment.
- If problem-gambling identification work is not carried through, DraftKings’ marketing and responsible-gambling functions operate with less internal counterweight to loss-focused targeting.
Second-order effects
- State regulators examining gambling safeguards gain a concrete basis to test whether operators’ promotional systems conflict with their consumer-protection commitments, extending the concerns raised by the earlier New York inquiry.
- Rival betting operators face pressure to explain how their own models separate customer-value scoring from problem-gambling safeguards, particularly where promotions are personalized.
Third-order effects
- The episode highlights an incentive conflict in data-driven gambling: the same behavioral data can optimize revenue from likely losers or identify risky play, making governance over model objectives a core compliance question.
- If regulators treat targeting logic as part of gambling oversight, operators may need auditable separation between acquisition models and responsible-gambling controls rather than relying on voluntary internal processes.
The trend: Sports-betting platforms are moving toward scrutiny of how machine-learning personalization allocates incentives and manages consumer-harm risk.