What ICT Challenges are being addressed in the ICT standardisation area?

As mentioned by the High Level Expert Group on Artificial Intelligence (AI), AI is one of the most transformative forces of our time and is bound to alter the fabric of society. Because of the challenges laying ahead, including societal acceptability and securing a strong industrial European industry in a global competitive market, it is of the utmost importance that AI can be trusted. Trustworthiness will come with the assurance that AI is technically robust and will include the ethics purpose that is founding our society, taking into account European values in contrast e.g. to Asian and US perspectives. The different elements of trustworthiness have to be thoroughly described and fully shared by all the stakeholders of AI. Therefore, standardisation is necessary.

While AI robustness is mainly a very technical issue and is therefore being tackled by the International Standardisation Organisation (through the work done in ISO/IEC JTC1/ SC42), the sensitivity towards the importance of taking into account values and principles remains very regional. In that view, Ethics purpose, while potentially an international societal concern, is a strong EU concern; which has to be prepared to address it as a future leader on ethics, other parties being less inclined to that topic.

How, if implemented will this make a difference in a specific context ?

As mentioned in previous paragraphs, Artificial Intelligence and Ethics are key to the transformation of our society. Without Ethics, AI might not be socialy accepted in some domains of our lives.

Are there any best practices that you are aware of that put into practice these challenges described ?

Artificial Intelligence and ethics are issues and challenges that are being addressed all around the world, the Ethics issue being a more European core value, even though it's being addressed as well in other parts of the world, but with less intensity. The ISO SC 42 is the international body working on AI. Ethics is partly addressed in ISO standard through the compliance and social responsibility standards. There is no known standardized best practice for Artificial Intelligence systems Ethics.

What future actions or further specifications work would at be necessary to undertake within an ICT Standards context?

The aim is, as chairman of the French AFNOR Artificial Intelligence standardisation committee:

• To participate to the ISO/SC42 international meetings
• To participate ti the EU CEN/CENELEC focus group on AI The future action will be identified through the work that will be done by the focus group