European AI Standards – Technical Standardisation and Implementation Challenges under the EU AI Act

Abstract
Harmonised standards are the cornerstone of efficient EU AI Act compliance. This paper presents one of the first systematic analyses of European technical and soon to be harmonised standardisation for organisations providing AI systems. Based on in-depth qualitative interviews with twenty-three leading European organisations developing AI applications across different sectors, such as Mistral and Helsing, and providing transparency regarding the status quo of draft standards, it examines how companies, especially start-ups and SMEs, are dealing with the contemplated standardisation under the EU AI Act and sectoral standardisation. Industry sectors covered include mobility, finance, manufacturing, healthcare, as well as defense and legal tech. Key challenges identified comprise an insufficient effective implementation period of likely less than 6 months compared to at least 12 months actually required for around thirty (partially referenced) technical standards, an imbalance of participation and influence in standardisation committees, double regulation and technical implementation hurdles as well as significant annual costs for harmonised standards compliance. Technical standards are currently reshaping global AI competition and will have a massive influence on the AI landscape as market entry barriers, particularly on start-ups. Hence, the paper offers policy recommendations based on the revealed challenges for AI providers.