Early indications are that artificial intelligence (AI) will have a transformative effect on player scouting within professional sport, including football. So long as elite sport remains competitive, clubs will find means to distinguish themselves from competitors. Where the economic reward for achievements such as domestic titles, qualifying for UEFA’s European competitions and avoiding relegation can be in the many millions of pounds financially, investment into the identification and employment of top talent is essential. AI-assisted scouting technology could be the “differential” clubs are looking for given the fine margins of success.

Where is AI being used in football?

The adoption of AI solutions by football clubs is not a new phenomenon. Google’s DeepMind has developed a generative AI tool to provide in-game tactical strategies that will improve teams’ set-piece productivity. Second Spectrum, a sports data company, uses machine learning technology to digitise live sport, creating accurate, informative data sets that could be used in real time decision-making, such as to signal an offside call or decide which team should have a throw-in. 

However, the opportunities presented through AI scouting are distinguished and arguably offer unique propositions, as clubs can add value to their academies and build bespoke squads with selected attributes, all at a significant time saving. 

AI scouting has been largely popularised by London-based technology company, ai.io, who have partnered with Major League Soccer (MLS) in the U.S. to globalise its “aiScout” app. The app provides partner clubs’ scouts data on the physical, technical and cognitive capabilities of targeted players, as well as the opportunity to view players performing videoed drills. In early 2024, LaLiga team Sevilla F.C. announced the launch of “Scout Advisor”, an AI tool for player scouting developed in collaboration with IBM. The tool implements Sevilla’s pre-existing player database and scouting network to assist in identifying and analysing potential signings. 

These instances demonstrate that the development and adoption of AI is presenting new opportunities for sports stakeholders to differentiate their club from competitors, but how does the technology at play differ between these disruptive scouting platforms?

AI Scouting Platforms and Technology

aiScout

aiScout is publicised as a democratisation of player scouting. It is built within a free mobile app, available to clubs and players alike. Interestingly, the app’s focus is on collecting data on amateur talent. Whilst professional clubs can typically access systematised data on professional players with relative ease, there is a deficiency of data available on players across amateur or lower, regional leagues, particularly for overseas leagues. aiScout currently contains data from roughly 100,000 players globally and is expected to increase this to 1 million by the end of 2024.

The utility of viewing recorded drills in the app is that players are benchmarked against other players, typically against players from the scout’s own team. aiScout leverages AI to provide concise data analytics and insights from these videos, automating ratings on various athlete metrics. This can serve as a useful pre-screening tool for clubs to ensure prospective signings meet their minimum requirements before engaging in traditional scouting methodologies.

Scout Advisor

Scout Advisor is marketed as an innovative AI tool which deploys both conventional and generative AI to enhance Sevilla’s scouting and player recruitment processes.

Scout Advisor utilises IBM’s 'watsonx' natural language processing (NLP) to search and evaluate the club’s scouting database. This includes searches and evaluations of both quantitative data, relating to player physique or statistics, and qualitative data, involving thousands of written scouting reports.

The platform applies multiple large language models (LLMs) to improve the accuracy and efficiency of player identification. Using prompts from Sevilla’s scouts, Scout Advisor can generate curated lists of players based upon the characteristics sought and summarises custom reports for each player. This streamlines a significant element of the traditional scouting process, freeing up time and resources for scouts to conduct deeper analyses on player recruitment and ultimately deliver better results on scouted players.

Intellectual Property and AI regulation in Football Scouting

Some of the key legal implications of AI leveraging sports data for player recruitment are centred around intellectual property and the regulation of AI. 

Intellectual property

The application of IP rights will differ across jurisdictions. However, in general, copyright and sui generis database rights are the IP rights relevant to the protection of databases, such as the databases of footballer performance data subject to AI analysis for scouting purposes. Copyright protects the arrangement of material in a database where this is original (i.e. creative) and sui generis database rights protect the content of a database where there has been substantial investment in obtaining, verifying and exhibiting the data, as has been held to be the case for live sports data in the UK and Europe. Importantly, copyright does not typically protect the underlying sports performance data (i.e. the database contents), only the “selection or arrangement of the contents” of the database.  It will therefore be relevant to determine how AI tools are sourcing the data sets they use to inform the scouts – i.e. is the data taken from pre-existing data sets (in which database rights are likely to exist) or does it create its own data sets based on analysis of players’ play?

Given the various rightsholders involved in professional sports, IP rights will typically be entrenched in commercial agreements driven by sporting bodies. It is important that clubs are aware of the underlying IP rights in the scouting database created and ensure there is appropriate provision made for the licensing of the relevant rights to avoid inadvertent infringement. 

Liability for AI program outputs remains a point of legal uncertainty which means, for the moment, clubs which engage AI scouting tools will need to interrogate how data sets are sourced and consider what their potential exposure to an IP claim might be. 

Regulation of AI

Given AI’s relatively recent surge and democratisation, AI regulation is still in its infancy. The EU AI Act, which came into force on 1 August 2024, aims to create a regulatory framework for the development and use of AI in the EU that ensures AI systems are secure, uphold current laws and adhere to general EU values. 

The Act takes a risk-based approach, delineating between the obligations of AI systems depending on their level of risk via the following profiles:

1. Unacceptable risk;
2. High risk;
3. Limited risk;
4. Minimal risk.

Such risk-profiling will be key to ensuring that AI systems are deployed safely, which is a key takeaway for football clubs with European scouting networks wanting to integrate AI solutions. Depending on risk categorisation, such clubs may be subject to material compliance requirements and potential enforcement action (monetary penalties and/or withdrawal of non-compliant AI systems). 

Conclusion

This evolution may bring an increased tension between the subjective assessment of scouts and the objective analysis of performance data in driving player recruitment. However, the creators and developers of aiScout and Scout Advisor affirm that these are ancillary AI platforms to enhance human scouting methodologies, rather than replace them.

We will continue to follow the development of use cases to see if these platforms turn from a secondary support system to primary drivers for player recruitment and, if so, how much clubs will be willing to invest to gain that extra margin and any key legal risks for clubs arising out of existing and future case studies.