EN 18286 becomes first harmonized standard under the EU AI Act

The reference to EN 18286:2026 was published in the EU’s Official Journal on 14 October, giving the standard legal significance, according to Techtidningen. The standard has been available since the summer and is the first harmonized standard under the AI Act.

EN 18286 covers quality management systems under Article 17 of the AI Act. It is aimed at providers of high-risk AI and describes how to organize quality work, including responsibilities, procedures and follow-up. Providers that follow the standard receive a presumption of conformity for the legal requirements it covers, according to CEN-CENELEC JTC 21.

Patric Ridell, project manager for the SIS AI committee, said the standard gives organizations a clearer way to show compliance. It does not provide general approval for an AI system: providers must still meet other applicable requirements, including those for risk management, data handling, technical documentation and cybersecurity. Using EN 18286 is voluntary; providers choosing another approach must show how it meets the requirements.

Ridell recommends that providers first inventory their AI systems, assess how the regulation applies and compare the standard’s requirements with existing processes. They should also decide how to document the work, assign responsibilities and follow up results. Quality work and monitoring must continue throughout an AI system’s lifecycle, including after launch.

Organizations using ISO/IEC 42001 do not need to start over, Ridell said. That standard has existed since 2023 and has a broader organizational focus; EN 18286 is tailored to the AI Act’s quality-management requirements for AI systems. The standards do not conflict, he said. EN 18286’s annexes also show links to ISO/IEC 42001 and ISO 9001.