AI Act compliance: Turning regulation into practical reality.

17-6-2025. The introduction of the European AI Act means that organisations working with AI-systems, or planning to do so, will be subject to a new legal framework. This framework adopts a risk-based approach: the level of obligations varies depending on the risk category of the AI-system and will be phased in over time. Since 2 February 2025, the provisions concerning prohibited AI-systems have come into effect. Additionally, from that date, organisations that develop, deploy, or use AI-systems are required to ensure an adequate level of AI literacy among their employees. For high-risk AI-systems, the obligations under the AI Act will be gradually enforced from August 2026 onwards. These obligations are stringent and often technically complex. This raises a key question: how can legal requirements be translated into concrete technical implementation? Harmonised standards play a central role in answering this question. These are technical standards developed at the request of the European Commission to clarify how compliance with the AI Act can be achieved.

In this blog, we outline the current developments around European AI standards, why they matter, and what this means for your organisation.

Current developments in European AI Standards (CEN and CENELEC)

To bridge the gap between the AI Act and the technical reality, the European Commission is relying on the instrument of harmonised standards. These standards are currently being developed by the Joint Technical Committee 21 (JTC 21), a joint initiative of CEN (European Committee for Standardization) and CENELEC (European Committee for Electrotechnical Standardization), two of the three European standardisation bodies responsible for developing technical standards across various sectors.The development of these harmonised standards is taking place under a formal standardisation mandate issued by the European Commission. JTC 21 is leading this comprehensive process, working closely with a wide range of stakeholders across Europe, including research institutions, regulators, and NGOs. The standardisation mandate explicitly requires the inclusion of diverse sectors and organisations to ensure the standards reflect a broad spectrum of needs, encourage innovation, and safeguard European values and interests. The standards are developed on a consensus basis. JTC 21 is leveraging existing international AI-standards, such as those from ISO and IEC, and adapting or supplementing them with new standards tailored to European requirements and supporting compliance with the AI Act. The ultimate goal is for these harmonised standards to serve as the practical implementation layer of the AI Act, providing developers of high-risk AI-systems with concrete tools for legal compliance. The complete set of standards is expected by the end of 2025 or possibly in 2026.

Why harmonised technical standards matter under the AI Act

Compliance with European harmonised standards, once they are published in the Official Journal of the European Union, creates a presumption of conformity with the AI Act. In other words, if an AI-system demonstrably meets these standards, it is presumed to comply with the legal requirements of the AI Act, unless proven otherwise. This mechanism offers legal certainty for companies and supports the creation of a level playing field across the European market. For small and medium-sized enterprises (SMEs) developing AI solutions, harmonised standards provide a more practical and accessible pathway to meet the complex requirements of the AI Act.

What does this mean for your organisation?

Implementing the AI Act requires a structured and proactive approach. Some provisions, such as the prohibition of unacceptable-risk AI-systems, have already been in force since February 2025, while the obligations for high-risk AI-systems will take effect starting August 2026. Organisations developing high-risk AI-systems should already be assessing the associated risks and implementing appropriate mitigation measures. Although harmonised standards will soon offer practical guidance, effective compliance begins with a thorough understanding of your organisation’s AI applications and their potential risks. Only by mapping the use and impact of your systems can you take targeted steps to meet the AI Act’s requirements and benefit from the support these standards provide.

What’s next?

At Considerati, we understand how complex it can be to comply with the AI Act. That’s why we support organisations with:

• Tailored AI literacy or AI governance training, focused on raising awareness of the legal, ethical, and organisational implications of AI.
• AI Inventory: a comprehensive mapping of all AI applications within your organisation, including risk assessment based on the AI Act.
• AI Gap Assessment: an in-depth analysis of your current processes, policies, and technical safeguards in relation to the AI Act’s requirements.
• AI Governance Framework: guidance in developing policy frameworks, responsibilities, procedures, and oversight mechanisms to manage AI use within your organisation.