Key data
| Regulation | Commission Implementing Decision (EU) 2026/2274 of 12 October 2026 |
|---|---|
| Official reference | OJ:L_202602274 |
| Publication in the Official Journal | 14 October 2026 |
| Entry into force | 12 October 2026 |
| Reference regulation | Regulation (EU) 2024/1689 — AI Act |
| Affected parties | Companies that develop, market or use high-risk AI systems in the EU |
| Main sectors | Technology, healthcare, transport and education |
| Category | European Regulation |
If your company develops, sells or integrates artificial intelligence systems in sectors such as healthcare, transport, education or technology, this regulation affects you directly. The European Commission has published, through the Implementing Decision (EU) 2026/2274, the harmonized technical standards applicable to high-risk AI systems under the framework of the Regulation (EU) 2024/1689, known as the AI Act. These standards, developed by European standardization bodies and referenced in the Official Journal, are the most direct way to demonstrate that your AI system complies with the law without having to build a technical justification from scratch.
What does this regulation establish?
The AI Act classifies artificial intelligence systems according to their risk level. High-risk systems are those that can significantly affect fundamental rights, safety or access to essential services. For these systems, the Regulation requires compliance with a series of technical and organizational requirements before they can be marketed or put into service in the European Union.
Implementing Decision 2026/2274 takes a key step: it publishes in the Official Journal the harmonized standards that European standardization bodies have developed specifically to address those requirements. The mechanism works as follows:
- A company that develops or markets a high-risk AI system can apply these harmonized standards in its design, development and documentation process.
- By doing so, it automatically obtains presumption of conformity with the corresponding AI Act requirements.
- This eliminates the need for additional technical justification before notified bodies or market surveillance authorities.
- The result is a reduction in the burden of proof and more agile access to the European single market.
Harmonized standards are not mandatory in themselves: a company can choose to demonstrate AI Act compliance through other means. However, following them is the safest, fastest and most cost-effective route to accredit conformity.
Economic and operational impact
The impact of this regulation translates into two areas: the cost of compliance and the cost of non-compliance.
Cost of compliance: Adopting harmonized standards involves investment in technical review, documentation and, in some cases, redesign of internal processes. However, this investment is significantly lower than the alternative: building your own technical justification before authorities, which requires considerable legal, technical and consulting resources.
Cost of non-compliance: The AI Act establishes significant sanctions for companies that market high-risk AI systems without meeting the requirements. The detailed summary of the regulation confirms that non-compliance can result in relevant economic consequences, although the exact amounts are set in Regulation (EU) 2024/1689 itself and not in this Implementing Decision.
Market opportunity: Companies that adopt these standards before their competitors will have a clear advantage: they will be able to accredit compliance more quickly, accelerate their certification processes and access the European single market earlier with their AI products.
| Scenario | Operational implication |
|---|---|
| Adopt harmonized standards | Presumption of conformity with the AI Act. No additional technical justification. Agile access to the EU market. |
| Demonstrate compliance through own means | Possible, but requires greater investment in technical and legal documentation. Higher risk of rejection. |
| Fail to comply with the AI Act | Risk of significant sanctions and blocked access to the European single market. |
Who does it affect?
This regulation affects any company operating in the high-risk AI value chain within the European Union. Specifically:
- Developers of AI systems classified as high-risk under the AI Act.
- Marketers and importers that place high-risk AI systems developed outside the EU on the European market.
- Technology sector companies that integrate AI into their products or services.
- Healthcare sector: diagnostic systems, clinical decision support, AI-based hospital management.
- Transport sector: autonomous driving systems, traffic management, infrastructure control.
- Education sector: automated assessment systems, access to training, academic guidance.
- CFOs and operations directors of companies that use high-risk AI in critical internal processes (personnel selection, credit scoring, etc.).
Practical example
A Spanish medical software company develops an AI system that analyzes diagnostic images to support clinical decision-making in hospitals. This system is classified as high-risk under the AI Act due to its potential impact on patient health.
Before the publication of harmonized standards, the company had to build its technical documentation from scratch to demonstrate to the notified body that it complied with the Regulation requirements. That process could involve months of technical-legal work and high consulting costs.
With Implementing Decision 2026/2274, the company can now apply the harmonized standards published in the Official Journal. By doing so, it automatically obtains presumption of conformity with the corresponding AI Act requirements, without need for additional justification. The certification process is streamlined, costs are reduced and access to the European market is accelerated.
What should companies do now?
- Identify whether your AI systems are high-risk according to the classification in Regulation (EU) 2024/1689. If you operate in healthcare, transport, education or critical technology, the probability is high.
- Review the harmonized standards published in Implementing Decision 2026/2274 and assess which ones apply to your specific systems.
- Update the technical documentation of your AI systems to reflect compliance with applicable harmonized standards.
- Involve the legal and technical teams in a joint analysis: harmonized standards have implications for both engineering and compliance.
- Plan conformity assessment processes with the corresponding notified bodies, taking advantage of the presumption of conformity granted by harmonized standards.
- Do not wait: the regulation has been in force since 12 October 2026. Each month of delay in adaptation is a month of exposure to sanctions and a month of competitive disadvantage compared to companies that are already certified.
Frequently asked questions
What is presumption of conformity with the AI Act and how is it obtained?
Presumption of conformity means that if a company applies the harmonized standards published in the Official Journal through Implementing Decision 2026/2274, it is automatically considered to comply with the requirements of Regulation (EU) 2024/1689 (AI Act) corresponding to those standards. It is not necessary to provide additional technical justification to market surveillance authorities. It is obtained simply by applying the harmonized standards in the design, development and documentation of the AI system.
Which sectors have high-risk AI systems according to the AI Act?
According to the regulation data, the sectors most affected by the high-risk classification are technology, healthcare, transport and education. In these areas, AI systems are more likely to impact fundamental rights, safety or access to essential services, which are the criteria that determine the high-risk classification under Regulation (EU) 2024/1689.
Is it mandatory to use harmonized standards to comply with the AI Act?
They are not mandatory. A company can demonstrate AI Act compliance through other technical means. However, applying the harmonized standards published in Decision 2026/2274 is the most direct and safe route: it grants automatic presumption of conformity, reduces the burden of proof and facilitates access to the European single market without need for additional technical justification.
When did this regulation come into force and when must companies be adapted?
Implementing Decision (EU) 2026/2274 came into force on 12 October 2026 and was published in the Official Journal on 14 October 2026. Companies with high-risk AI systems must review their compliance status immediately, as the regulation is in force and non-compliance with the AI Act can result in significant sanctions.
What happens if a company markets a high-risk AI system without complying with the AI Act?
Regulation (EU) 2024/1689 establishes significant sanctions for companies that fail to meet its requirements. In addition to the economic risk from sanctions, companies may face blocked access to the European single market for their AI products. Implementing Decision 2026/2274 precisely facilitates compliance by publishing the harmonized standards that grant presumption of conformity.
Official source
Consult full regulation at official source
Disclaimer: This article is for informational purposes only and does not constitute legal advice. For specific decisions, consult a qualified professional. Source: https://eur-lex.europa.eu/./legal-content/AUTO/?uri=OJ:L_202602274