AI evaluates footballers. EU rules raise new questions about scouting

22 October 2026

Norwegian football clubs already use artificial intelligence to analyse and evaluate players. Meanwhile, the EU is setting new requirements for AI systems used in recruitment and candidate assessment. For Norwegian sportstech, this raises a question: when is AI-based scouting sporting analysis, and when might it become part of regulated recruitment?

Norwegian football clubs already use artificial intelligence to analyse and evaluate players. Meanwhile, the EU is setting new requirements for AI systems used in recruitment and candidate assessment. For Norwegian sportstech, this raises a question: when is AI-based scouting sporting analysis, and when might it become part of regulated recruitment? News analysis. When Norwegian Top Football (NTF) <a href="https://www.eliteserien.no/nyheter/norsk-toppfotball-styrker-klubbene-med-banebrytende-norsk-teknologi">announced its partnership with Fokus on 18 December 2024</a>, the message was smarter player recruitment and squad planning. The Norwegian start-up had developed a tool using AI to identify and evaluate players. According to the press release, Bodø/Glimt had used the solution since spring 2022. AI in Norwegian scouting is therefore not just a future prospect. The technology is already used in work that can lead to a club signing a player. But there is an important distinction between supporting that work and making the contractual decision. The Fokus agreement documents data-driven scouting and player evaluation. It does not document automated contract decisions or a particular legal classification. With the EU’s new high-risk rules, the product’s place in the decision chain becomes a commercial question for suppliers. <h2>From passing analysis to candidate assessment</h2> AI can help clubs process large datasets, compare performances and identify players for closer assessment. That does not make every such function a recruitment system. An analysis of passing quality or movement patterns may serve a different purpose from recommending who should receive a professional contract. The latter may bring a system within the scope of <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/annex-3">Annex III, point 4 of the AI Act</a>. This provision includes systems intended for recruiting or selecting people, analysing and filtering applications and evaluating candidates. Certain decisions and performance assessments in employment relationships are also covered. “Performance analysis” is therefore not, by itself, a legal safe harbour. A hypothetical product ranking candidates for professional player contracts needs a different assessment from a tool describing match sequences. What matters is the system’s intended purpose and specific function, not whether its supplier calls it “scouting” or “analysis”. This example does not determine Fokus’s classification. <h2>High risk also concerns decision support</h2> Following the <a href="https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32026R1744">amendments in Regulation (EU) 2026/1744</a>, the relevant high-risk rules for systems covered by Annex III apply from <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-113">2 December 2027 in the EU</a>. This does not mean every scouting tool becomes high-risk on that date. <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-6">Article 6</a> sets classification criteria and allows exceptions, including certain narrow or preparatory tasks where the system does not pose a significant risk. Its influence on the outcome of decisions requires assessment. For systems falling within Annex III, however, profiling natural persons must always be considered high-risk. A provider claiming an Annex III system is exempt must document its assessment. A sporting director making the final decision does not automatically provide an exemption. A system can influence which candidates are seen, compared or excluded before the meeting where a contract is discussed. This potential influence makes decision support relevant, without bringing every statistical comparison within scope. High-risk classification can entail requirements for <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-9">risk management</a>, <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-10">data quality</a> and <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-11">technical documentation</a>. Clubs must receive <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-13">information enabling them to understand outputs</a>, and systems must facilitate <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-14">effective human oversight</a>. Oversight means being able to understand limitations, override outputs and avoid excessive reliance on the model. <h2>A digital development pathway is not automated selection</h2> NFF’s partnership with Kitman Labs illustrates another part of this development. The <a href="https://www.kitmanlabs.com/news/norway-football-federation-selects-kitman-labs-to-unify-national-player-pathway/">announcement on 11 June 2026</a> describes a shared platform intended to connect around 1,900 club nominations, 18 football districts and the youth national teams. The solution will support Landslagsskolen, Norway’s national player development model, through nominations, training-camp administration, coaches’ evaluations and longitudinal tracking, among other functions. The announcement describes infrastructure for sharing information and following players’ development. It does not document NFF using AI for automated ranking or selection of young players. Digital registration and monitoring, AI-based performance evaluation, algorithmic selection for development programmes and recruitment to professional employment contracts are different functions. A public agreement covering the first does not prove the other three. Selection for a youth national team is not automatically employment recruitment under Annex III either. For young athletes, privacy, accurate information and fair evaluation nevertheless matter. Access to player data, the purpose of collection and the ability to correct errors must be addressed independently of high-risk classification. These are general considerations in talent development, not documented shortcomings at NFF or Kitman Labs. <h2>The product’s role may determine market access</h2> Norwegian suppliers must distinguish the EU deadline from Norwegian implementation. The <a href="https://www.regjeringen.no/no/aktuelt/tung-vil-sende-ki-loven-med-endringer-pa-horing/id3169693/">government stated on 4 August</a> that work on EEA incorporation and a Norwegian AI law was ongoing, with the ambition of presenting legislation in spring 2027. The announcement is not a Norwegian commencement date. Norwegian businesses may nevertheless fall within the <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-2">EU regulation’s scope</a> when placing systems on the EU market or where their outputs are used in the EU. Product, role and use require individual assessment. A Norwegian company registration does not, by itself, shield an export product. For sportstech companies, this makes defining functionality an early strategic choice. An analysis tool and a product intended to rank contract candidates may use related data but face different requirements. Documentation, human oversight and contractual allocation of responsibilities between supplier and club should develop alongside the functions, rather than being added after a sale. For clubs and investors, it becomes relevant to understand what the product is actually intended to do, its limitations and how the supplier can substantiate its reliability. Responsible AI can build trust and support European market access. The strategic challenge is not just how sophisticated the analysis becomes, but the role it plays in decisions affecting players’ careers. Illustration of footballers on a training pitch at sunset. The image does not show the systems discussed in use.