Safe and Trustworthy Artificial Intelligence Standards: Commission Notifies Intention to Adopt Standardisation Request

Stakeholders have until 3 January 2023 to submit feedback by email to the draft Commission Implementing Decision on a standardisation request to CEN / CENELEC  in support of safe and trustworthy artificial intelligence. This Implementing Decision aims to prepare the necessary technical ground for the implementation of the future AI Act. The text of the draft measure includes the following:

Article 1 - Requested Standardisation Activities and Requirements

The draft Implementing Decision would aim to ensure the safety and trustworthiness of AI in the EU.

The Commission would request CEN and CENELEC to draft the new European standards or European standardisation deliverables on the following requirements – specified in annex II:

1. risk management system for AI systems;
2. governance and quality of datasets used to build AI systems;
3. record keeping through logging capabilities by AI systems;
4. transparency and information provisions to the users of AI systems;
5. human oversight of AI systems;
6. accuracy specifications for AI systems;
7. robustness specifications for AI systems;
8. cybersecurity specifications for AI systems;
9. quality management system for providers of AI systems, including post-market monitoring process; and
10. conformity assessments for AI systems.

The deadline for their adoption by CEN and CENELEC would be set for 31 January 2025.

Article 2 - Work Programme

CEN and CENELEC would be required to prepare a work programme indicating all requested standards, the responsible technical bodies, as well as the timetable for the execution of the requested standardisation activities, in line with the deadlines laid down under Annex I. The European Telecommunications Standards Institute (ETSI) should be consulted during the preparation of the text.

In addition, CEN and CENELEC would be required to submit the work programme to the Commission by four months after the Commission’s adoption of the standardisation request and provide the Commission access to an overall project plan.

Article 3 - Reporting 

CEN and CENELEC would be required to report every six months to the Commission on the execution of the standardisation request and indicate the progress made in the implementation of the work programme.

Moreover, a joint semestrial report should be submitted to the Commission ten months after the adoption of the standardisation request, while a joint final report should be submitted by 31 January 2025.

Article 4 - Validity of the Standardisation Request

Should the standardisation organisations not accept the request referred to in Article 1, the request would not constitute a basis for standardisation activities.

Lastly, the standardisation request would expire on 30 November 2025.

Interplay with the EU AI regulatory landscape

The expected Commission Implementing Decision was first announced in the Commission’s 2022 Annual Union work programme for European standardisation.

The measure would request CEN and CENECT to draft harmonised standards to: (i) lay down rules for the placing on the market, putting into service, and use of AI systems in the EU; and (ii) address requirements related to their safety and trustworthiness, including risk management, data quality, transparency, human oversight, accuracy, robustness and cybersecurity.

The objective of the measure would be to ensure the safety and trustworthiness of AI systems, that such systems are properly monitored throughout their lifecycle, that they respect fundamental values and human rights recognised in the EU, and that they strengthen European competitiveness. The measure making reference to the Artificial Intelligence Act.