A joint statement issued on Monday announced the creation of the “GPAI”, a global alliance for the regulation of artificial intelligence (AI).
The group includes the EU (with several member states also signing up individually), the US, the UK, Australia, Canada, India, Japan, Mexico, New Zealand, South Korea and Singapore. Its core aim is to support the responsible development and use of AI. The focus of the group will be to ensure that AI respects human rights and “shared democratic values”, as explored in the OECD's Recommendation on AI. It is clear from the joint statement that the GPAI will work closely with the OECD in achieving its goals.
The GPAI intends to establish a working group in each of the four key areas that it has identified:
- responsible AI;
- data governance;
- the future of work;
- innovation and commercialisation.
Within each area, GPAI aims to bring together leading experts from across society (public and private sectors) to collaborate on cutting-edge research. Early signs are encouraging: reports suggest that Yoshua Bengio, a recent Turing Award winner, will co-chair the working group on Responsible AI.
The GPAI will be supported by two centres - in Montréal and Paris - which will support the Working Groups. They will arrange an annual GPAI Plenary, the first of which will be hosted in Canada in December 2020. Highlighting the close co-operation with the OECD, the GPAI will also be supported by a Secretariat within the OECD in Paris.
Whilst the GPAI has a long-term focus, it acknowledges the importance of the fight against COVID-19 and will investigate how AI can be used in the global response. AI has already played a key part, from being used in the development of contact tracking apps and in the analysis of huge datasets (in context of outbreak monitoring and in assisting medical researchers, for example) to the detection of signs of COVID-19 in patients (using tools such as algorithms to analyse CT images).
In the context of increasing calls for international collaboration on how best to regulate AI, including explicit support from big private players, the creation of the GPAI should be seen as welcome development. Progress is slow, but this is a move in the right direction.