The European Commission has cited EN 18286:2026, Artificial intelligence: Quality management system for EU AI Act regulatory purposes, in the Official Journal of the European Union through Commission Implementing Decision (EU) 2026/2274, published on 14 October 2026. It is the first harmonised standard to be cited under the AI Act (Regulation (EU) 2024/1689). For manufacturers of medical devices and IVDs that contain a high-risk AI system, it gives a recognised way to meet the AI Act’s quality management requirements.
What the citation does
Under Article 40 of the AI Act, a high-risk AI system that conforms with a harmonised standard cited in the Official Journal is presumed to conform with the requirements that the standard covers. EN 18286 addresses the quality management system that providers of high-risk AI systems must put in place under Article 17. Its scope includes design and development controls, risk management, data management, records and post-market monitoring.
Three limits are worth keeping in mind:
- Using the standard is voluntary and a provider can show compliance with Article 17 by other means (unlikely).
- The presumption only extends to the requirements the standard covers. It does not cover the other obligations for high-risk AI systems, such as the technical requirements on data governance, accuracy, robustness and cybersecurity.
- The standard concerns the provider’s management system and does not evaluate the performance of a particular model or AI system.
Before the citation, EN 18286 was a CEN and CENELEC publication that manufacturers could use as a reference. After it, the standard is the benchmark against which an AI Act quality management system is likely to be assessed.
Why it matters for medical devices and IVDs
AI systems that are, or are safety components of, devices requiring notified body involvement under the MDR or IVDR are classified as high risk under Article 6(1) of the AI Act. Following the AI Act Omnibus, the obligations for these Annex I products apply from 2 August 2028, while the Annex III high-risk obligations apply from 2 December 2027.
Device manufacturers already operate a quality management system, usually based on EN ISO 13485:2016, to meet Article 10 of the MDR or IVDR. Article 17(3) of the AI Act allows providers that are subject to quality management obligations under sectoral Union law to include the AI Act elements in that existing system. Manufacturers do not need to build a second, separate quality management system for the AI Act.
EN 18286 helps with that integration. It sets out what the AI Act expects of a quality management system in a form that can be compared, clause by clause, with an existing ISO 13485 system. JTC 21 describes informative annexes that map the standard to EN ISO 9001 and EN ISO/IEC 42001, and mentions no mapping to ISO 13485, so device manufacturers should expect to prepare that mapping themselves.
Where we expect gaps
From the structure of the AI Act and the content of EN 18286, the areas where an ISO 13485 system is most likely to need additions are:
- data management for training, validation and test data, including data governance and records of data provenance
- risk management that covers risks to health, safety and fundamental rights, alongside the ISO 14971 process
- post-market monitoring of AI system performance, including performance drift
- change control for AI models, including how retraining and updates are assessed
- records and logging that support traceability of the AI system’s operation
What to do now
There is time before the obligations apply, and that time is best used for planning.
- Confirm whether each software product or AI component in your portfolio is a high-risk AI system under Article 6(1) of the AI Act.
- Obtain EN 18286:2026 and map its requirements against your ISO 13485 quality management system.
- Record the gaps and decide how each will be closed within your existing procedures.
- Discuss with your notified body how it intends to assess AI Act quality management requirements as part of MDR or IVDR conformity assessment.
- Follow the further harmonised standards being developed by CEN and CENELEC JTC 21 on risk management, data governance, logging, transparency, human oversight, accuracy, robustness and cybersecurity.
Specculo supports manufacturers of software and AI as a medical device with regulatory strategy, IEC 62304 lifecycle work, ISO 13485 quality systems and the EU AI Act. If you would like help mapping EN 18286 against your quality management system, get in touch.