Standardization for Compliance in the European Union’s AI Act
The European Union’s Artificial Intelligence Act (AI Act) aims to establish a comprehensive framework for the regulation of AI technologies. A crucial element of this framework is the role of standardization, which is expected to provide technical solutions to ensure compliance with the Act’s complex requirements.
The Importance of Standardization
Standardization is seen as a key mechanism for achieving compliance with the AI Act. As various stakeholders engage in the development of relevant standards, the challenge remains significant: standards must be ready for implementation by August 2026, coinciding with the majority of the AI Act’s provisions coming into effect. The European Committee for Standardization and the European Electrotechnical Committee for Standardization are tasked with making these standards available by the end of 2025.
Standards and Specifications
Within the AI Act, several types of standards are defined:
- Harmonized Standards: AI systems classified as high-risk and complying with harmonized standards published in the Official Journal of the EU will be presumed to meet the requirements outlined in the AI Act. These standards also encompass the general transparency requirements under Article 50 of the Act.
- Common Specifications: The Commission has the authority to adopt common specifications for high-risk and limited-risk requirements. However, harmonized standards take precedence, and common specifications can only be adopted if no suitable standards exist.
Certificates and Compliance Assessments
Providers of high-risk AI systems must ensure their products undergo a conformity assessment procedure before entering the European market. This procedure varies based on the system’s type, and it can involve either internal control or an external review by a Notified Body (NB).
Upon successful assessment, an NB will issue an EU technical documentation assessment certificate, which is crucial for demonstrating compliance with the AI Act. These systems must also bear the CE marking to signify their conformity.
Role of Notifying Authorities and Notified Bodies
The AI Act mandates EU Member States to designate at least one Notifying Authority (NA) responsible for overseeing the assessment and notification of conformity assessment bodies (CABs). NAs must ensure that CABs can conduct independent and objective evaluations of high-risk AI systems.
Notified Bodies play a vital role in this ecosystem, verifying the conformity of AI systems according to the established assessment procedures. They must meet rigorous requirements to ensure impartiality and confidentiality in their assessments.
Conclusion
As the EU moves forward with the implementation of the AI Act, the development of standards and specifications will be crucial in meeting regulatory requirements. Stakeholders must engage in collaborative efforts to ensure that these standards are not only developed but also effectively integrated into the compliance frameworks of AI technologies.
This comprehensive approach aims to foster innovation while safeguarding public interests, thus balancing the advancement of AI technologies with necessary regulatory oversight.
