DIGITALEUROPE recommendations on standardisation in the field of Artificial Intelligence
Standards can play an important role in complementing emerging policies, laws and regulation around AI by connecting objectives and requirements with practical implementation. Standards also contribute to fostering innovation and enabling interoperability.
EU institutions should promote engagement in AI global standardisation initiatives and leverage global standards development to support EU AI policies and regulation.
Initial AI standardisation work should focus on finding consensus around AI foundational concepts, management and governance practices.
AI standards need to be applicable to a variety of contexts and should not constitute barriers to trade.
Compliance with policies and regulatory requirements should be achieved without mandating specific technology choices.
Artificial Intelligence (AI) is a major opportunity for innovation and growth in Europe. Many papers and recommendations have been issued by European organisations calling for standardisation to raise trust and thereby promote the adoption of AI technologies and solutions.
Indeed, trust in new technologies, such as AI, and innovation around these technologies is best supported when policy objectives and regulatory requirements make use of voluntary industry-driven standardisation to support implementation. That way, compliance with policies and regulatory requirements as well as interoperability between different implementations can be achieved without limiting the potential for innovation by mandating specific technology choices.
At the same time, we caution against inappropriate use of standardisation. AI is global. While nations and regions may compete for AI innovation power, their industries need global scale to effectively compete and develop their activities. That is best achieved through global standards rather than regional and national ones.
For the same reason, standards are not an appropriate tool for codifying cultural norms or values: AI standards can, and should support adherence to ethical principles but they need to be adaptable to a wide variety of value systems.
AI standards should also not be used such that they establish barriers to trade. The EU institutions should encourage engagement in global WTO TBT-compliant standardisation initiatives.
The immediate challenges that standardisation should address are:
Establishing consensus around AI foundational concepts, management and governance practices.
Framing concepts and best practices to establish trustworthiness of AI, including in areas such as privacy, cybersecurity, safety, reliability, and transparency.
Global standardisation efforts in these areas are ongoing, for instance within ISO/IEC JTC 1 Information Technology and the IEEE Standards Association. It is crucial that European institutions promote the awareness of this work amongst member states, such that the international standards developed take into consideration any EU-specific policy objectives relating to AI.
Additionally, the EU Rolling Plan for ICT Standardisation is a useful tool to assist with awareness raising of AI standardisation work already underway. It is important that the European Commission continues to engage, via the Multi-Stakeholders Platform on ICT Standardisation, in the annual Rolling Plan updating process and use it to specify standard-related action requests, which can be taken up by the standards developing community.