EN 18286:2026, a quality management system standard for AI providers, has been published in the Official Journal of the EU, becoming the first harmonised standard under the AI Act. It was published through Commission Implementing Decision 2026/2274 of 12 October, on 14 October, World Standards Day. The standard maps to the Article 17 requirements for a quality management system. Providers of high-risk AI systems that apply it gain a presumption of conformity with those requirements.

What it means: Harmonised standards are the practical way companies show they meet the general requirements of an EU regulation. If a provider of a high-risk AI system applies EN 18286, it is presumed to meet the quality management system obligations, without having to prove each point on its own. This is an important step, because delays in standards were one of the main arguments for postponing the high-risk obligations. One standard is not enough, though, since other requirements such as risk management, data quality, logging and human oversight still lack published standards. For Greek software companies building AI applications in areas such as recruitment, credit scoring or health, the standard gives a clear guide to organising their processes. Those already certified to ISO 9001 or ISO/IEC 42001 will find much in common. Good documentation now will ease both conformity assessment and relationships with public sector and financial clients, who will ask for proof of compliance.

Read the original on CEN-CENELEC

Related service: myDATA integrations

Related stories