What does fAIr play look like: AI and Gaming in 2025

By Caitlin Vanderkarr, Heidi McNeil Staudenmaier, and Tony Caldwell

Sports betting. AI. Both subjects evoke plenty of attention across the world. When AI migrates into the realm of sports betting, a new level of curiosity emerges.

The term “AI” refers to artificial intelligence, which involves a “computer system[]… perform[ing] complex tasks normally done by human-reasoning, decision making, creating.”¹ AI has been so pervasive in the past several years that even the United States federal government created several executive orders describing its own use of AI.² Generative AI is a subset of AI which uses data to create outputs like text, photos, and even “love poems”.³

Sports betting, or “gambling that involves wagering on sports”⁴ has been legalized in several jurisdictions within the U.S. since the United States Supreme Court struck down the Professional and Amateur Sports Protection Act (PASPA) in 2018.⁵ Sports betting occurs both online and in-person, although the lion’s share occurs online. In 2023, online sports betting represented over $10 billion out of $11 billion total sports betting revenue.⁶

Given that sports betting and AI are both prevalent online, their combination is not surprising. Sports betting AI tools can be used to develop strategy, analyze performance, anticipate and formulate odds algorithms, and much more.⁷ Forbes identified several AI tools used in sports betting, including for the purpose of player performance analysis.⁸ For example: Sportradar Fraud Detection System (FDS) (used for integrity purposes to analyze betting patterns and flag irregular activity); PlayScan (used to identify risky gambling behavior and provide alerts and resources regarding problem gambling); BetBuddy (used to assess individual betting styles and histories to make predictions tailored to a risk profile and preferences); and Stratagem and Stats Perform (used to analyze historical data, player statistics, and weather).⁹ eSports also utilizes AI for insights and predictions, including Unikrn, as applied to Dota 2 or League of Legends.¹⁰ Not one to be left behind, the PGA Tour plans to launch “TOURCAST” so viewers can “follow flight paths and gain access to… analytics for every shot.”¹¹

Most recently, FanDuel implemented “AceAI” for a small pool of its patrons to use with NFL and NBA wagers.¹² AceAI is a form of generative AI that can converse with FanDuel patrons.¹³ FanDuel’s VP of Product Transformation and Innovation, Jon Sadow, stated that FanDuel “ha[s] a tremendous amount of data that [it] can actually put into customers’ hands that empower them to make educated decisions,” which includes acting as a virtual assistant to “assemble … parlays,” “assist[ing] in detecting warning signs of problem gambling… [and] direct[ing] customers to a variety of resources” for problem gambling, and answering historical wagering inquiries.¹⁴

So, why should the average sports bettor care about AI?

AI may save time, analyze data extremely quickly to allow a data-based decision, and/or create a level of interactivity that may not be reasonable in terms of human capital required to implement such interaction.

Is there a business case for AI within gaming?

The short answer, from industry leaders, seems to be yes. Industry-wide AI utilization is one of the top cited key changes anticipated for sports betting in 2025.¹⁵

DraftKings CEO Jason Robins voiced enthusiasm for AI, even referencing a discussion with his co-founder contemplating a future where humans “manag[e] teams of AI”, but indicating that the right “policies and procedures” are key to safe implementation.¹⁶ DraftKings is currently seeking regulatory approval for a “sports predictions market”— a trading market style “‘contract[]’ for a team to win an event” that amounts to a wager against other bettors in the trading market — that, if legal, perhaps might integrate the type of customer-facing AI tool Robins mentions.¹⁷

However, no discussion about sports betting can truly be complete without analyzing the sports themselves that are the subject of such wagers. 76ers’ President Daryl Morey indicated in a discussion on the “Pablo Torre Finds Out” video podcast that his team uses “AI or machine learning to forecast draft picks… [with] traditional and linear regressions.”¹⁸ In reference to artificial general intelligence (AGI) in this same discussion, MIT Economics and EECS Departments Professor, author, and non-profit co-founder Sendhil Mullainathan, envisions a space where “algorithms… do things that people could never do” rather than replace human contributions, because “we already have people who can think like people.”¹⁹ With sports teams implementing AI in this way, it is conceivable that sports betting could fully integrate AI, from sports teams’ AI-assisted athlete selection to sport betting platforms’ backend technology, and even as a tool for sports betting patrons to assess their wagers before placing them.

One publicly available AI source, ChatGPT, has captivated the public since its creation. OpenAI, the company that created ChatGPT and previous iterations of large language models, maintains publicly available application programming interfaces (APIs) on its website so developers can integrate ChatGPT (or OpenAI’s other large language models) into other applications.²⁰ ChatGPT has been used to create sports outcome predictions. Gary Pearson published an article that enumerated ChatGPT’s picks for the Australian Grand Prix, including the rationales shared by the AI engine.²¹

Given the aforementioned use cases, it appears that AI’s influence and integration into gaming may continue to expand from multiple angles. Savvy folks in the gaming industry would do well to explore AI and proactively focus on ethical, regulatory, and policy issues that may arise from AI’s inclusion in the gaming space. AI can be a powerful ally, not only to enhance gaming experiences for bettors, but to facilitate financial monitoring (including “transaction histories, detecting potential money laundering activities, and flagging suspicious transactions”).²²

Governmental entities are currently developing AI laws and regulations, especially within sports betting (a heavily-regulated industry itself). Within the U.S., as of 2024 “there [wa]s no comprehensive federal legislation or regulation” regarding AI, but there were over 120 bills before Congress regarding AI.²³ Even this may change within 2025, as the FTC may be stepping to the forefront as a potential AI regulatory body and several states have enacted their own AI-related legislation (including Colorado and California in 2024 with respective enforcement dates of 2026 and 2025).²⁴

Meanwhile, the EU has the “AI Act” (2024); China, India, the United Kingdom, and other countries are developing their AI laws currently; and the Global Partnership on Artificial Intelligence (GPAI) formed a collaborative effort between over 40 countries regarding AI.²⁵  The EU’s AI Act, with a recent implementation date of February 2, 2025, requires “prohibited AI practices [] be withdrawn from the market… includ[ing] real-time biometric surveillance tools, systems that subliminally manipulate people’s behavior[,] and AI that makes automated decisions based on protected characteristics like race or gender.”²⁶ EU AI regulation is expected in May, with compliance deadlines anticipated in August of this year; this regulation and enforcement schema applies to both large and small businesses.²⁷ Canada, Japan, Australia, and the G7 countries are all considering next steps for AI.²⁸

So, from a business perspective, how might a company or organization prepare for AI in sports betting?

ChatGPT’s response to this question mirrors several concepts already mentioned by tech leaders:

AI input aside, there are a few actionable items a business could proactively undertake now:

AI will likely continue to grow. Sports betting industry insiders should stay informed, adapt quickly, and plan ahead. Another idea: engage with AI now (in an ethical and appropriate manner) to better understand its limitations and capabilities.

[A note of disclosure: AI did not draft this article, but, in the spirit of innovation, a version of AI was used to verify bluebook citation format (a time saving approach) and ChatGPT was consulted on one question (pure curiosity).]