On September 23, 2021, ITI hosted the virtual event, “Global Perspectives on Artificial Intelligence Policy: How to Further Global Cooperation?”
In case you missed it, you can find a recap of the event below or watch the event here:
ITI brought together global policy makers and industry experts to discuss perspectives on artificial intelligence (AI) policy and potential for global collaboration. The event featured distinguished panelists from major global economies: Juha Heikkilä, Adviser for Artificial Intelligence, European Commission, José Gontijo, Director of the Department of Science, Technology and Digital Innovation at the Brazilian Ministry of Science, Technology and Innovation, Yoichi Iida, Deputy Director General for G7 and G20 Relations, Global Strategy Bureau, and the Japanese Ministry of Internal Affairs and Communications, and Elham Tabassi, Chief of Staff of the Information Technology Laboratory at the U.S. National Institute of Standards and Technology.
Panelists agreed that while the take up of AI technologies has incredible potential, it is also important to take into account potential risks that it carries. Elham Tabassi stressed in this regard the need to develop standards and increase trust on AI to fully reap the benefits of the technologies, while José Gontijo and Yoichi Iida both remarked how global approaches to AI policy will help support global innovation and economic growth, as well as establishing guidelines for potential regulation around this technology. Yoichi Iida also stressed the value of implementing AI principles based on human centricity and democratic values in a balanced and measured fashion.
“We all know that developing technology standards development is an increasingly an area of global strategic and economic competition. International standards help overcome trade barriers and expand market access, helping [...] support global economic growth.” - Elham Tabassi, U.S. National Institute of Standards and Technology
The debate then shifted to how testing and regulation could be enforced across jurisdictions. Juha Heikkilä and José Gontijo both used the example of the European GDPR and the adoption of similar measures by other countries to illustrate what international harmonization on AI measures may look like. Yoichi Iida agreed, adding that interoperability, communication, and mutual understanding between countries was very important. Elham Tabassi echoed Yoichi Iida’s sentiments by outlining the US National Institute of Standards’ actions to work towards a global understanding and interoperability around AI use and risk management. Panelists agreed that there was no one-size-fits-all approach, but that interoperability was crucial.
“In the area of AI regulation, I personally feel that different countries have to have different frameworks because we have different social requirements. But on the other hand, we need interoperability between these different frameworks.” - Yoichi Iida, Ministry of Internal Affairs and Communications, Japan
The categorization of high-risk AI cases and potential for global convergence in this regard was then brought to the table. José Gontijo shared Brazil’s approach, explaining how any definition of high-risk AI needs not to hinder innovation in the field. Panelists also discussed that high-risk AI should be based on the use case rather than on the technology itself, and international forums such as the OECD or G20 can play a role to establish a consensus on which uses should be treated as such. Juha Heikkilä stressed that risks to safety and fundamental rights can be the starting point for a common definition of high-risk AI between like-minded jurisdictions.
“Product safety and fundamental right risks could be seen as the basis for a high-risk AI definition in other jurisdictions who take a similar human-centric approach to AI .” - Juha Heikkilä, European Commission
During the Q&A, panelists discussed potential next steps to encourage international collaboration and avoid divergence. Elham Tabassi stressed the importance of moving from principles to practice by developing standards and metrics for AI risk-management, while Juha Heikkilä and José Gontijo agreed that cooperation between all parties involved would be a key factor in achieving results and building consensus. Yoichi Iida strongly supported these comments, stressing the need to closely collaborate among like-minded countries towards a human-centric use of AI.
“We need to involve our national AI ecosystems in global conversations to support the global development of the technology and input on the development of public policies at home.” - José Gontijo, Brazilian Ministry of Science, Technology and Innovation
To stay up-to-date on ITI work in Europe, including future events, subscribe to our ITI Europe newsletter.