DraftKings Is Using AI to Supercharge the Harms of Online Behavioral Advertising
Original reporting by Electronic Frontier Foundation
A recent investigation has revealed how the online sports betting company DraftKings is deploying artificial intelligence to identify and target customers most likely to place losing bets. By training machine learning models on users’ betting records, DraftKings pinpoints individuals deemed profitable due to their propensity for losses. These customers then receive personalized advertisements designed to entice them back to the platform, capitalizing on their vulnerability—a practice particularly harmful to those identified as “problem gamblers.”
This predatory approach exemplifies the magnified harms of online behavioral advertising when supercharged by AI. While data-driven advertising has long personalized content, the integration of AI amplifies its effectiveness, incentivizing the collection of even vaster quantities of personal data. AI's "black box" nature further drives this data hunger, as developers seek ever more information to refine opaque models. The consequences extend beyond individual exploitation; the immense datasets gathered for ad targeting inadvertently fuel a broader surveillance industry, with data finding its way to insurers, banks, and government agencies. DraftKings’ use of solely first-party data in this scenario underscores a critical policy challenge: simply limiting third-party data sharing is insufficient. Organizations like the EFF argue that the only effective solution to mitigate these harms is a comprehensive ban on all online behavioral advertising, thereby removing the core incentive for such intrusive data collection.
The DraftKings situation serves as a stark illustration of how AI, when integrated into existing predatory advertising models, can amplify harm and exploit human vulnerabilities. By leveraging sophisticated machine learning on first-party data, companies can pinpoint individuals most susceptible to their offers, turning what might be a recreational activity into a targeted trap for those with problem gambling tendencies. This immediate consequence — the direct exploitation of financially and psychologically vulnerable customers for profit — underscores a critical ethical breach in the application of AI.
The Broader Landscape
Beyond the immediate ethical concerns, DraftKings’ approach highlights a broader, more insidious shift in the digital landscape. The relentless pursuit of user data to fuel increasingly precise AI models not only personalizes ads but also solidifies the foundation of a pervasive surveillance economy. This constant data ingestion, often opaque and operating within "black box" algorithms, enables companies to extract value from user behavior in ways that transcend simple product recommendations. The implications extend far beyond online betting; this same data-driven, AI-powered exploitation model could easily be repurposed to target individuals based on financial distress, health conditions, or political leanings. Future policy solutions, therefore, must move beyond merely regulating third-party data to address the fundamental incentive for data collection itself, advocating for a ban on all behavioral advertising to truly safeguard individual privacy and autonomy in an AI-driven world.
Frequently asked questions
- How is DraftKings using AI and customer data to influence online sports betting behavior?
- DraftKings employs AI to analyze customer betting records and identify individuals prone to making losing bets. The company then targets these users with personalized promotions designed to encourage continued engagement and further wagers. This strategy aims to maximize profit by re-engaging customers who frequently lose, unfortunately including those who may be problem gamblers, rather than mitigating their risk of harm.
- How does AI enhance the harms associated with online behavioral advertising practices?
- AI magnifies the harms of online behavioral advertising by enabling faster processing of vast data sets and more precise targeting. Its "black box" nature incentivizes continuous data collection, as developers seek to understand model efficacy. This leads to an unprecedented accumulation of personal information, which can be exploited not only for predatory advertising but also potentially sold to other entities like insurance companies, banks, or government agencies.
- What are the wider privacy implications of data collected for online behavioral advertising?
- Data collected for online behavioral advertising has privacy implications extending beyond ad targeting. This sensitive personal information is often sold to various third parties, including insurance companies, banks, and even state and federal law enforcement agencies like CBP and ICE. This practice fuels a broader surveillance industry, raising significant concerns about how personal data, gathered initially for commercial purposes, can be used for other, often undisclosed, activities.