Sports Q&A: How is AI being utilised in Sport?

The proliferation of AI technology and the availability of vast amounts of data have revolutionised, and will no doubt further transform, the use of data in sports. By using computer systems and algorithms to perform tasks that would normally require human intelligence (e.g. machine learning, problem-solving, and decision making), AI is being used to analyse vast amounts of data, provide real-time feedback, and enhance the overall performance of athletes and teams.

In this Q&A we discuss some of the key questions on the use of AI in Sport.

Q. What are the key AI concepts and terms relevant to the sports industry?

A. Computer vision – a subfield of AI that enables computers to capture image and video data (using cameras) and interpret and analyse the information received by applying machine learning models to identify and classify objects, altogether simulating the way humans see and understand their environment by replicating the human eye.

Machine Learning - an application of AI that can automatically learn and improve from experience without being explicitly programmed to do so. AI which implements machine learning operates in a similar way to a computer code which carries out statistical analysis: in processing an enormous amount of data, it looks for patterns and suggests a better understanding of that data to inform future use.

Datasets – a collection of various types of data stored in a digital format and a key component of any machine learning project.

Q. How is AI being used in a sports context?

A. AI is used in a multitude of different ways in a modern sporting context, including:

• Injury Prevention - Coaches are combining AI and video to predict injuries before they happen. Injury prevention AI technology works by analysing large amounts of data, including player movement (via computer vision) and health data, to identify potential risks of injury. The technology uses machine learning algorithms to analyse this data and identify patterns that could lead to injuries. This analysis is then used to create personalised training programs for athletes to mitigate the risks and prevent injuries. Major sporting teams have already started to deploy such technology with Liverpool FC using Zone7’s technology since the start of the 2021/2022 season.
• Performance Analysis - Similarly, computer vision and wearable technology is being used to monitor and analyse real time performance and data from athletes in order to provide insights and recommendations to coaches. This technology can be used to identify areas of weakness or opportunities for improvement during a game. A prime example would be Second Spectrum – which analyses performances using advanced computer vision and machine learning algorithms to track the movements of players in the NBA and produce reports for coaches to analyse and identify patterns in a player or a team performance.
• Scouting and Recruitment – The use of data science in the recruitment and scouting of athletes has only improved with the use AI as scouts can track player attributes, value, skillsets and future potential. AiSCOUT allows amateur players to upload videos of themselves performing drills set by a prospective club and their performance is then compared to the performance by players of the prospective club completing the same drills using computer vision analysis, allowing scouts to access a new talent pool and benchmark performance against their current players. AiSCOUT’s platform has already been successfully deployed by notable football clubs, including Chelsea, Burnley, Nottingham Forrest and Olympiacos.
• Fan Engagement – AI provides the opportunity for fans to receive further insights and data analytics on their favourite teams and players. In the same way performance analysis can help coaches, the use of computer vision creates performance data for fans to enjoy. F1 insights powered by AWS uses around 300 sensors on each F1 vehicle generating data points which provide real-time race predictions and car performance data, creating insights that “inform, educate, and enrich the fan experience”.

Like other businesses, sports organisations and businesses also have the opportunity to use AI in non-sports specific scenarios. We’ve explored some of these opportunities, and the risks that come with them, in other articles, see for example: AI 101: How do AI tools work and why are lawsuits being raised? and AI 101: What are the infringement risks of using AI-generated works.

Q. What regulation covers the use of sports AI systems?

A. Draft EU AI Regulation seeks to harmonise rules on artificial intelligence across the EU by ensuring AI products are sufficiently safe and robust before they enter the market. The AI Regulation is intended to apply to what the EU terms (in the current text of its draft AI Regulation) as “AI systems” - all software developed through machine learning approaches, logic and knowledge-based approaches and statistical approaches. This will cover most AI models deployed in a sport context. The European Parliament has recently proposed a new definition to future-proof the scope of the draft AI Regulation by moving away from defining an AI system as “software” and instead defining it as a “machine-based system that is designed to operate with varying levels of autonomy and that can, for explicit or implicit objectives, generate outputs such as predictions, recommendations, or decisions that influence physical or virtual environments”. As of the date of publication, these proposals have not yet been adopted.

The EU AI Regulation takes a ‘risk-based approach’ to governing AI systems which conform with the defined criteria. Not all AI systems will be subject to the obligations under the AI Regulation. The AI Regulation divides AI systems into different strands of risk, based on the intended use of the system. The many ways in which sporting organisations utilise their AI systems are likely to fall under the high-risk AI system category (e.g. scouting and recruitment).

The UK Government recently outlined its contrasting proposed legislative framework in its ‘White Paper on Artificial Intelligence’. The White Paper sets out proposals for implementing a "proportionate, future-proof and pro-innovation framework for regulating AI", striking a balance between protecting public interests and fostering innovation, focussing on establishing a flexible regulatory system which evolves alongside technological advancements. This approach should ensure that businesses in the UK, including sport businesses utilising AI models, can continue to develop and deploy AI technologies without being held back by overly restrictive regulations.

Q. What are the data protection concerns related to using AI in sports ?

A. AI models used by sports teams rely on vast datasets to not only operate, but also to continuously learn and improve. It is likely that much of this data collated by AI systems will consist of personal data of players and/or fans. When processing personal data, organisations using AI models must comply with relevant data protection legislation (e.g. the UK/EU GDPR), which sets out certain obligations. For example, the data should be used in a fair, lawful and transparent manner, meaning that organisations will need to establish a lawful basis for using the data (likely consent or establishing a legitimate interest) and ensure that affected individuals are appropriately informed. Data should also be used for clearly defined purposes and appropriate security measures need to be taken to ensure confidentiality, integrity and availability of the data. Data accuracy is also important, and there is a question as to how AI can identify and rectify mistakes.

Aside from the general principles described above, specific rules around the processing of ‘special categories’ of personal data (including health data) are relevant to the tracking of player performance and injury. There are additional requirements for processing such data (it can only be used if certain conditions are met) and heightened security measures should be in place to ensure adequate protection.

Use of special categories of data has been challenged by data subjects in the sports industry; players from England’s and Scotland’s football leagues joined together in 2020 to bring claims against gambling and data aggregation companies, arguing that their performance data had been used without their consent. Other sporting organisations, including (recently) the Professional Cricketers’ Association, have joined the “Project Red Card” movement, which seeks legal redress (and compensation) for the profits that gambling and data processing companies have enjoyed. Last year, letters before action were sent to the relevant companies, meaning that legal proceedings may take place in the future if players’ grievances are not settled outside of court.

The use of AI systems to make decisions about athletes (e.g. using AI systems for team selections) also raises data privacy issues. Subject to some limited exceptions, Article 22 of the UK GDPR gives data subjects the right not to be subject to solely automated decisions (i.e. the making of decisions by automated means without any human involvement), including profiling, which have a legal or similarly significant effect on them. These provisions restrict when a processor can carry out this type of processing and gives data subjects specific rights in those cases.

Q. Who owns the IP rights in datasets?

Broadly, there are two types of intellectual property protection for databases (and individual datasets): copyright and sui generis database rights. These rights allow the owner to control certain uses of their databases. Copyright protects the selection or arrangement of material in a database where this is original (i.e. creative) and sui generis database rights protects the content of a database where there has been substantial investment in obtaining, verifying or presenting the data. Sports institutions may face challenges when relying on the sui generis database right to protect the data they create because this right safeguards the investment in obtaining, verifying, or presenting the data rather than the data creation itself.

The application and operation of the copyright and sui generis database rights, including their territorial scope, are beyond the scope of this Q&A, however, at a high-level, the ownership of the data in a sports context will ultimately come down to the commercial arrangements regarding the data. These arrangements will delineate which stakeholders own the datasets, if any. Third-party tech providers, AI model developers, sports teams, sports leagues and individual players will negotiate agreements setting out the terms and conditions of use, and establish a clear ownership structure for the data, so that stakeholders can exploit data without coming across constraints from any other party.

As a side note, from a privacy perspective, where personal data is shared between organisations, the extent to which a data recipient can exploit personal data for its own purposes will depend on its commercial arrangements with the supplier of the data and its ability to establish that it can use such data in a manner that complies with data protection legislation.