The Robots Are Taking Over: The Use of Artificial Intelligence in the NFL and Solutions to Combat Uncertain Legal Landscape

By Kaitlyn Furst*                                                                             Posted: 04/09/2024

On February 11, 2024, about 123.7 million viewers watched Super Bowl LVIII, which featured an exciting rematch between the San Francisco 49ers and the Kansas City Chiefs.[1]  Although Super Bowl LVIII set television record history, it also made technological history as its programming featured the NFL’s debut of its Digital Athlete program and an unprecedented number of generative artificial intelligence (AI) advertisements. [2]  Many of these commercials featuring generative AI focused on the technology and how consumers could use it. [3]  Besides permitting AI advertisements during its programming, the NFL has also embraced this technology as a tool to predict and prevent player injuries. [4]  Although the NFL is using AI technology to benefit its players and staff and increase fan engagement, the NFL must be conscious of the legal ramifications that AI might bring upon the organization, and thus, adopt policies that comply with both United States privacy and copyright laws.[5] 

What is AI?

AI technology has been used for years despite generative AI programs becoming popular among the public in 2023. [6]  Traditionally, AI models have been used to make predictions based on data; however, in recent years, generative AI has been designed to create new original creative content, such as images, video, computer code, text, and audio [7]  As a result, these generative AI programs have widely manufactured “more objects that look like the data it was trained on.”[8]  Specifically, programs such as ChatGPT and OpenAI are widely accessible to the public, thus creating the likelihood that the way certain jobs are performed will change in many industries. [9]

NFL’s Use of AI Technology with Players

During the 2023-2024 football season, the National Football League (NFL) announced its partnership with Amazon Web Services to create an injury and health prevention tool, known as “Digital Athlete.”[10]  Digital Athlete utilizes both AI and machine learning (ML) to create interactive and realistic gameplay stimulations, allowing NFL teams to analyze players’ game time experiences and health information to determine which players are at the highest risk for injuries. [11]  For its realistic simulations, Digital Athlete gathers data from a multitude of sources including video filmed during games and practices, game day data from the Next Generation Stats System, information from stadium sensors, and player tracking data from “tracking devices embedded in players’ equipment.” [12]  The NFL has touted that this “technology is ‘the first’ of its kind” and hopefully will allow athlete trainers to adapt appropriate training methods for each respective player. [13]

Potential Legal Ramifications of NFL’s Current AI Use

Although the NFL’s cutting-edge use of AI is intended to benefit players, teams, and fans, there are certain legal ramifications with using this technology, specifically regarding the impact of both privacy and publicity rights and copyright concerns.[14]  Currently, there are no federal regulations for AI regarding privacy protections; however, some states, such as Pennsylvania and New York, have proposed legislation to combat problems with AI-generated deepfakes.[15]  While there are no federal AI protections yet, there is currently federal legislation pending imposing stricter data privacy requirements on American companies.[16]  It is likely that if this legislation is passed, that it could potentially be extended to AI use by American companies or set the foundation for substantial federal AI legislation. [17]

Regardless of whether there is federal AI legislation in place, there is a concern that the disclosure of player health information is not enumerated in the NFL’s Collective Bargaining Agreement (CBA). [18]  Specifically, the CBA provides that player health information will only be used for general research involving NFL players, not player-specific information. [19]  Even if the NFL reserved its right in the CBA to adopt a different policy or process for using this data, this provision appears to solely focus on procedures regarding the NFL’s Injury Surveillance System, and therefore, does not state whether this applies to newly developed programs by the NFL.[20]

Besides considering concerns involving the language of the CBA, the NFL is also bound by applicable copyright law, which could potentially affect the AI material shown during advertisements on their programming.[21]  Although the Copyright Act does not explicitly define “author,” currently, the U.S. Copyright Office only recognizes creative works by a human. [22]  In fact, the U.S. Copyright Office has put out guidance indicating that this authorship requirement does not apply to generative AI programs, unless a human edited any portion of the material, potentially creating issues for the NFL if an advertiser edits an AI‑generated advertisement without their knowledge. [23]  Moreover, even if an advertisement is completely AI-generated, there are still concerns indicating “contractual restrictions imposed by AI tool providers” regarding “attribution obligations or commercial use limitations.” [24]

Suggestions

Considering all the uncertainty in the legal landscape regarding AI programs, the NFL should strongly consider adopting policies regarding AI use to combat any potential legal issues in the future.[25]  The laws regarding AI are in flux, and there are many individuals and groups in the United States advocating for protections to be adopted at the federal level to combat these issues on a larger scale. [26]  As a result, since the NFL is using AI more than other sports leagues, they must adopt policies that require advertisers to disclose that advertisements were solely created by generative AI programs or were edited by a human. [27]  Besides protecting the NFL from defending against any copyright infringement claims, this would also promote transparency regarding the use of AI in any broadcasting, thus, strengthening the NFL’s relationship with its fans. [28]

Additionally, if the NFL wants to adopt a program such as Digital Athlete to be used by all its teams, then the NFL should make sure its players are aware of the health information that the program is capturing through the sensors on their equipment. [29]  While it would be beneficial for this information to be explicitly included in the CBA, considering this CBA is in place until 2030, it is important that the NFL explicitly send out information in writing about this program and the data it collects, who has access to this information, how the information is stored, and how this information is being used. [30]  The NFL should adopt a similar process to the medical authorizations and require that each player sign that they understand this program and allow an opportunity for athletes to opt out of this program if they are not comfortable with how the information is being collected or used. [31]  As a result, implementing a policy like this will also promote transparency between the NFL and its players, thus strengthening support for their digital athlete program.[32]

*Staff Writer, Jeffrey S. Moorad Sports Law Journal, J.D. Candidate, May 2025, Villanova University Charles Widger School of Law

[1] See Joe Reedy, Super Bowl was the Most-Watched Program Ever in the US, Averaging 123.7 Million Viewers, AP (Feb. 13, 2024, 3:42 PM), https://apnews.com/article/super-bowl-ratings-2024-chiefs-49ers-cbs-e2f0288ea474bd564ee4df430e41b8bd (stating that Super Bowl LVIII was not only most watched Super Bowl game, but also “most-watched program in U.S. television history.”); see also Super Bowl LVIII Draws 123.7 Million Average Viewers, Largest TV Audience on Record, Nielsen (Feb. 2024), https://www.nielsen.com/news-center/2024/super-bowl-lviii-draws-123-7-million-average-viewers-largest-tv-audience-on-record/#:~:text=NEW%20YORK%20%E2%80%93%20February%2013%2C%202024,single%2Dnetwork%20telecast%20to%20date (acknowledging there was partnership between “NFL, CBS, Paramount+, Nickelodeon and Univision” to increase audience engagement, including “Spanish-language broadcast on Univision” and “kid-friendly broadcast on Nickelodeon and Nick-at-Nite.”).

[2] See Michelle Castillo, The Way AI Will be Sold to a Skeptical Public is Starting to Become Clear, CNBC (Feb. 15, 2024, 10:09 AM), https://www.cnbc.com/2024/02/13/the-way-ai-will-be-sold-to-a-skeptical-public-is-becoming-clear.html (implying artificial intelligence companies used Super Bowl LVIII as platform to “humanize” artificial intelligence technology).

[3] See PYMNTS, AI Took Over this Year’s Super Bowl Ads; It’s Coming for Daily Life Next, PYMNTS (Feb. 12, 2024), https://www.pymnts.com/news/artificial-intelligence/2024/ai-took-over-super-bowl-ads-daily-life-is-next/ (illustrating Super Bowl LVII commercials that featured generative AI targeted consumers to show how AI could be integrated into one’s life to improve “many facets of daily life and commerce.”).  Advertisements during Super Bowl LVII showed that AI could be used to protect against cyber fraud with identity theft and security breaches.  See id. (contrasting AI commercials during 2024 Super Bowl to 2022 Super Bowl as commercials during 2024 Super Bowl were focused on showing ways AI technology could prevent with data issues, while 2022 Super Bowl commercials were focused on using AI technology to make cryptocurrency investments); see also Elizabeth Napolitano, Microsoft’s Super Bowl Message: We’re an AI Company Now, CBS News (Feb. 9, 2024, 3:20 P.M.), https://www.cbsnews.com/news/super-bowl-2024-microsoft-ad-commercial-ai-copilot/ (explaining Microsoft’s Super Bowl LVII advertisement emphasized that they are now integrating AI technology into their company, ultimately assisting individuals with generating computer code and creating digital art).  Microsoft’s advertisement that featured AI focused that this technology could assist with making daily decisions, as well as more creative decisions such as finishing movies, designing video games, and developing study plans.  See PYMNTS, supra note 3 (explaining ways Microsoft used their new AI technology, AI Copilot, in their Super Bowl LVII advertisement).

[4] See Kayla Bailey, NFL Revolutionizes Sports by Using AI to Prevent Injuries: ‘It Will Have a Profound Impact,’ SVP Says, Fox Bus. (Feb. 11, 2024, 7:00 AM), https://www.foxbusiness.com/sports/nfl-revolutionizes-sports-using-ai-prevent-injuries-will-profound-impact (explaining that purpose of using AI in this manner is to “help trainers work in a more ‘efficient but smarter fashion’ so that they can better” adapt to each player’s needs).

[5] For further discussion regarding solutions to assist NFL with AI issues, see supra notes 26-32 and accompanying text.

[6] See Adam Zewe, Explained: Generative AI, MIT News (Nov. 9, 2023), https://news.mit.edu/2023/explained-generative-ai-1109 (explaining AI in previous years referred to learning models that only made predictions depending on preexisting data).

[7] See id. (explaining differences between traditional AI programs and generative AI, focusing new research on building larger datasets to develop new models that can develop groundbreaking creations).  Traditional AI programs were based in a statistical method to determine the data output, which is like systems that have autocomplete functions.  See id. (illustrating that traditional AI technologies were less complex in their functions).  Generative AI programs are built on machine learning algorithms, which use a larger dataset that can result in many different data outputs and does not have as much predictability as traditional AI programs.  See id. (emphasizing generative AI programs focus more on data patterns, allowing it to predict what should come next instead of autocompleting data set).

[8] See id. (indicating how generative AI platforms are based on “machine-learning model” and that this model uses both research and mathematical data based on formulas from at least fifty years ago).

[9] See Francesca Lazzeri, Generative AI, OpenAI, and ChatGPT: What are They?, Medium (May 30, 2023), https://medium.com/data-science-at-microsoft/generative-ai-openai-and-chatgpt-what-are-they-3c80397062c4 (implying generative AI could become commonplace in industries as its use has “more than doubled” in last five years and many companies are actively investing in this technology).

[10] See Jennifer Langton, Building a Digital Athlete: Using AI to Rewrite the Playbook on NFL Player Safety, NFL Player Health & Safety (Feb. 1, 2024, 12:00 PM), https://www.nfl.com/playerhealthandsafety/equipment-and-innovation/aws-partnership/building-a-digital-athlete-using-ai-to-rewrite-the-playbook-on-nfl-player-safety (explaining new partnership between NFL and Amazon Web Services to create platform that allows NFL teams to recognize needs of their individual players so they can stay healthy and perform at their best).

[11] See id. (acknowledging data collected is used to assist NFL teams in understanding needs of their individual players to play at their best and stay healthy).  Besides using AI and ML to predict player injuries, Digital Athlete is used to run simulations of various game plays, thus creating a “safety playbook” and informing any potential safety initiatives by the NFL.  See id. (explaining Digital Athlete creates simulations of game play by using “sets of player, weather, equipment, and stadium data”).

[12] See id. (summarizing Digital Athlete’s sources of data so it can run realistic simulations for NFL teams that illustrates player experience during specific games).  For example, the game time simulations specifically use location, speed, and acceleration information from various data sources to make sure a realistic simulation is produced.  See id. (explaining Digital Athlete can produce accurate player experience of games, which can allow teams to create millions of simulations so they can be better informed of injury risks during specific games).  An important feature of Digital Athlete is its ability to recognize helmet impacts, which could ultimately determine the amount of force that could cause a concussion.  See id. (acknowledging that Digital Athlete currently measures helmet impacts, but Digital Athlete is currently being modified to measure helmet forces to assist in concussion prevention).

[13] See Bailey, supra note 4 (implying that if Digital Athlete is successful, this program will have “profound impact” on sports and possibly could be used in other industries such as military).

[14] See Shirin Malkani & John Delaney, AI at Super Bowl Raises Contract, Copyright Issues for Pro Sports, BL (Feb. 9, 2024, 4:30 AM), https://news.bloomberglaw.com/us-law-week/ai-at-super-bowl-raises-contract-copyright-issues-for-pro-sports (arguing AI technology imposes contractual limitations on both use and engagement of AI technology and could create further issues with team’s players based on terms of collective bargaining agreement).

[15] See id. (explaining there are no federal protections regarding AI generated material, but it is likely that state legislation will continue as use by public and popularity of these tools increase).

[16] See American Data Privacy and Protection Act, H.R. 8152, 117th Cong. (2022) (explaining that this bill would establish requirements for American companies for how they handle customers’ personal data).  These requirements would force companies to both limit collecting this data and using the personal data when providing a product or service and would require a customer to consent to a company transferring any identifiable data.  See id. (summarizing changes to companies’ data privacy requirements and that these restrictions would be enforced by Federal Trade Commission and state attorneys general).

[17] See Cameron F. Kerry, How Privacy Legislation Can Help Address AI, Brookings (July 7, 2023), https://www.brookings.edu/articles/how-privacy-legislation-can-help-address-ai/#:~:text=The%20ADPPA%20contains%20several%20important,or%20for%20other%20enumerated%20purposes (arguing provisions in American Data Privacy and Protection Act about algorithmic accountability and fairness could create “baseline for AI regulation.”).  Since the American Data Privacy and Protection Act would create restrictions on “[t]he collection, use, and sharing of personal information,” it is likely that these restrictions will need to be used for AI technology since this technology is using personal data in both its analysis and output materials.  See id. (acknowledging that AI technology’s power comes from its ability to use personal information and that these restrictions in American Data Privacy and Protection Act should be used as framework for future AI protections).

[18] See generally 2020 NFL-NFLPA Collective Bargaining Agreement, Art. 39, §§ 1-21 (2020), available at https://nflpaweb.blob.core.windows.net/media/Default/NFLPA/CBA2020/NFL-NFLPA_CBA_March_5_2020.pdf (illustrating that use of player information or any personally identifiable information would be used by NFL on platform designed by third-party).

[19] See id. at Art. 39, § 18 (explaining that IQVIA will be using health information to produce injury/illness reports and other research to adopt new policies and programs to promote health and safety); see also id. at Art. 40 (summarizing procedures for medical records and health information, indicating that this information is protected by signed medical record authorizations).

[20] See id. at Art. 39, § 18(a) (summarizing that injury/illness research and analysis is completed by IQVIA and various advising committees, panels, and boards, but it does not indicate whether this broad provision is applicable to collecting data not from medical records).

[21] See Copyright Act, 17 U.S.C. §§ 101-810 (2022) (illustrating any claim for copyright is protected under this act, which protects “original works of authorship fixed in any tangible medium of expression”).

[22] See id. (illustrating that authorship is not explicitly defined, and thus, requires the interpretation of the Courts or the U.S. Copyright Office); see also Naruto v. Slater, 888 F.3d 418, 425-26 (9th Cir. 2018) (holding that animal, explicitly monkey in this case, lacked statutory standing under Copyright Act since it was not human).

[23] See Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16190-01 (proposed Mar. 16, 2023) (to be codified at 37 C.F.R. pt. 202) (indicating that artificial technology is not considered author or co-author if human only uses generated material produced by generative AI program and does not modify it in any way); See Carl A. Kukkonen III, Emily J. Tait, & Joshua H. Gold-Quiros, Court Finds AI-Generated Work Not Copyrightable for Failure to Meet ‘Human Authorship’ Requirement – But Questions Remain, Jones Day (Aug. 2023), https://www.jonesday.com/en/insights/2023/08/court-finds-aigenerated-work-not-copyrightable-for-failure-to-meet-human-authorship-requirementbut-questions-remain (arguing decision in Thaler v. Perlmutter created issue of “just how much human participation will be required . . . to meet human authorship requirement”).  Thaler held that only humans qualify as authors and that AI-generated work could not be eligible for copyright protection unless there are modifications to the work by a human.  See id. (indicating that Thaler created question of how much human involvement is required to meet human authorship requirement if work is majorly AI-generated content).  It is likely the question of how much human authorship is required could become an issue in copyright infringement cases.  See id. (acknowledging there could be issues as to when generative AI materials could be infringing on third-party works and whether human authorship could affect this outcome).

[24] See Malkani & Delaney, supra note 14 (arguing sports leagues broadcasting games could require approval rights regarding use and engagement of AI technology or require AI disclosure to protect themselves from potential issues regarding AI material).

[25] For further discussion on potential legal ramifications of AI use by the NFL, see supra notes 14-24 and accompanying text.

[26] See Hayden Field, AI Lobbying Spikes 185% as Calls for Regulation Surge, CNBC (Feb. 2, 2024, 12:01 P.M.), https://www.cnbc.com/2024/02/02/ai-lobbying-spikes-nearly-200percent-as-calls-for-regulation-surge.html#:~:text=Artificial%20intelligence%2Drelated%20lobbying%20reached,OpenSecrets%20on%20behalf%20of%20CNBC (illustrating that many Americans are advocating for AI regulation and that Biden Administration is also pushing to codify these protections at federal level).

[27] See Kukkonen, et al., supra note 23 (implying that any individual applying for copyright protection must disclose any solely AI-generated content in work being submitted for registration). Although advertisers do not submit their advertisements for copyright, they must disclose this information so that their advertisements do not result in a claim for copyright infringement against the NFL. See id. (arguing that there needs to be transparency regarding these AI-generated works).

[28] See Sandy Carter, AWS, AI, and Taylor Swift Usher in the Super Bowl Show Sports’ Future, Forbes (Feb. 10, 2024, 7:00 A.M.), https://www.forbes.com/sites/digital-assets/2024/02/10/aws-ai-taylor-swift-and-usher-in-the-super-bowl-show-sports-future/?sh=30000b34a5bf (illustrating that use of AI in Super Bowl made it more interactive and personalized, creating new fan experience that many fans did not recognize until after end of broadcast).

[29] See Langton, supra note 10 (acknowledging amount of data collected on players for this program run by both NFL and Amazon Web Services).

[30] See 2020 NFL-NFLPA CBA: Need to Know, NFL Football Operations, https://operations.nfl.com/inside-football-ops/players-legends/2020-nfl-nflpa-cba-need-to-know/ (last visited March 24, 2024) (indicating 2020 CBA is in place through 2030 season); see also Player Health & Safety, NFL Football Operations, https://operations.nfl.com/inside-football-ops/players-legends/player-health-safety/ (last visited March 24, 2024) (emphasizing importance of NFL and NFLPA working together to protect player health and players know what is happening with medical information).

[31] See 2020 NFL-NFLPA Collective Bargaining Agreement, Art. 39, Appendix X available at https://nflpaweb.blob.core.windows.net/media/Default/NFLPA/CBA2020/NFL-NFLPA_CBA_March_5_2020.pdf (explaining authorizations and protocols required for “dissemination and use” of player health information). Appendix X requires NFL medical teams to consent to various protocols regarding any player injury data or related information. See id. (emphasizing that any request for injury data or related information must follow all protocols in Appendix X and that NFL medical teams must understand these protocols). Currently, based on the language in Appendix X, only NFL medical teams, including all team physicians and athlete trainers, must understand and consent to this information being released. See id. (summarizing procedures for released NFL player medical data depending on type of information requested and what organizations or individuals will have access to this information).

[32] See Molly Smith, The NFL Must Prioritize Players’ Safety and Humanity, Chi. Pol’y Rev. (Jan. 19, 2023), https://chicagopolicyreview.org/2023/01/19/the-nfl-must-prioritize-players-safety-and-humanity/ (arguing NFL “must be transparent about the health risks associated with [football]” and importance of reporting); see also Langton, supra note 10 (indicating that Digital Athlete program will assist in building trust regarding injury prevention).