The “The Age of AI – Sports and Sports Betting” Webinar offered a brief overview of how generative AI (“GenAI”) works and how AI is being used or is planning to be used in the sports and sports betting industries, and finished with identifying the key challenges raised by GenAI and relevant considerations when selecting vendors and vetting and implementing GenAI tools.
It’s likely that many employees within sports and sports betting companies are already using or experimenting with ChatGPT or other GenAI tools, and that management has already taken at least initial steps to guide employee usage or has already released formal policies, while also preparing for the changing risks and challenges ahead stemming from GenAI use.
Commenting on some GenAI output examples, the presenters illustrated that sometimes GenAI tools miss the nuance of a user’s prompt or, for example, might not comprehend sports lingo, which is why an organization would generally not rely on public-facing GAI tools in the long-term but instead might consider acquiring a license for an enterprise version that allows the AI model to train on the company’s data and content so that it can create more effective and accurate output.
The presenters offered several reasons why sports and sports betting companies might use GenAI, such as to increase productivity of routine tasks, generate raw ideas more quickly, personalize experiences for users in real time, and drive consumer engagement and retention (thereby helping companies stay competitive). To be sure, the benefits of GAI must be balanced with the risks, which include the potential for inaccurate or slanted output based on the algorithmic bias of the AI model’s training materials, the possibility of offensive or unexpected outputs to consumer users (if GenAI is deployed in a customer-facing interface), and the costs and investments that must be made to customize an AI application (e.g., hardware, time and expertise, employee training, fine-tuning and cleaning of data, uploading proprietary data, cybersecurity).
The presenters then described some AI-related use cases in the industry, which might involve usage in marketing (social media engagement; personalized ads); media and user experience (creation of customized content, automated highlight reels, immersive experiences, user-generated GenAI art based on your company’s own brand and content); analytics (real-time data integration, enhanced betting odds, management of athlete performance); and partnerships (AI-related sponsorships and licensing opportunities).
Overall, the presenters stressed that AI-generated content can add value to sports, marketing and fan engagement. Some specific ways include:
- AI tools can scan video libraries and quickly create highlight reels, even personalized reels that match a fan’s favorite players or team (or even a fantasy sports’ participant’s players).
- GenAI can accelerate the production time for in-game content that can be reactive to what fans are talking about. Similar advances might aid in generating real-time statistics.
- GenAI can generate sports commentary, albeit without personality or emotion. The presenters cited the use of “Henry” from IBM’s AI Commentator that provided commentary on the Masters App to fill in the silence of shot videos that lacked sports commentary. The model was trained to convert tracking data into language descriptions.
- GenAI can be combined with augmented reality to enable immersive experiences, such as giving fans the virtual feeling of dunking a basketball (e.g., in the NBA App, fans can scan themselves, create an avatar, and select a player in a live game and replace it with this avatar and watch the game with the avatar making the moves of the actual player).
- Social media is an important part of some fans’ experience (the “second screen”), and GenAI tools can analyze high-interest subject matter and measure consumer sentiment – helping to leverage this information in fan engagement strategies and deliver more relevant content to your audience. Still, the presenters noted that it is important for the marketing team to have a human involvement in this process, whether it be editing or valuing the GenAI tool’s suggestions.
- Sports betting is regulated on state-level basis and the presenters noted the availability of GenAI-powered tools to answer compliance questions about the changing U.S. regulatory landscape in sports betting.
- Conversational AI, if properly trained, can help with responses to customer’s needs.
- GenAI can help sportsbooks with predicative analytics. Sportsbooks are already using a suite of software and technology to run operations, along with confidential algorithms that analyze third party data and pricing feeds to produce effective betting odds. GenAI can be used to optimize odds even further and could be used to collect and analyze more complex sports data (e.g., more precise movement tracking of the players on the field) and betting markets.
- GenAI can be used to evaluate player performance and support scouting, salary negotiations, player acquisition, and simulate drafts and trade deadlines to allow management to be better prepared to make decisions. GenAI can also inform game decisions in real-time based on conditions and other variables. In the future, AI might also be used to predict athlete injury, thus aiding in planning out rest and training schedules. Advanced systems are already being employed to aid in refereeing, line calls and umpiring.
The presenters finished the Webinar by offering some key considerations to evaluate which use cases makes sense and how to implement GenAI in the most responsible way. The presenters stressed that confidentiality of your proprietary or sensitive data should be one of the paramount considerations when deciding which GenAI tools to use and what sorts of corporate policies to develop. If an employee is uploading sensitive data or code into a GenAI tool, that is a potential disclosure (a fact that may not be self-evident to employees), thus making public-facing GAI tools riskier to use for businesses due to this confidentiality risk, as compared with enterprise arrangements that are more tailored to the business environment. The presenters stressed that current rules on confidentiality should apply equally to new GenAI tools. Another important issue is thinking about brand protection and develop policies around how others might use GenAI tools to create content that uses your IP. Of course, the use of GenAI tools may implicate publicity and privacy rights, requiring necessary compliance.
Lastly, businesses planning to use custom GenAI applications must comprehend the level of resources required to commit to these types of initiatives and ensure the return on investment is enough. The presenters discussed many important questions to ask, including: What is your strategy on GenAI and how do you want to use it now and in the future? What are the policies surrounding GenAI usage? What person or group will oversee these decisions and inform employees about these shared goals related to the technology? Similarly, the presenters stressed the importance of assessing potential partners: What is the degree of customization required for a particular GenAI tool? Does it need to be trained on your own proprietary data sets or can it simply be trained on general industry or public domain data, and what additional personnel or expertise is required to upload and structure the data for processing by the AI model? Does the vendor offer enterprise options that protect your data behind “walled gardens” so proprietary data does not become part of the vendor’s training data? As with every vendor relationship, it is important to mitigate risk with an appropriately worded agreement.
