Re-engineering the systems for effective governance of AI

The multi-stakeholder panel of experts entitled "Re-engineering the systems: best practices and building blocks for effective governance of AI" focused on emerging best practices in the supervision of AI. UNESCO's Assistant Director-General for Social and Human Sciences, Gabriela Ramos, emphasized that implementation is key to adapting governance structures to steer AI for good, highlighting its urgency and significance. 

According to a 2023 report from the World Privacy Forum, a review of 18 AI governance tools used by governments found that more than a third (38%) include problematic or ineffective tools to evaluate and measure AI systems, particularly for fairness and explainability. It shows the need for more effective governance frameworks. During the panel, experts shared their assessment of how to make AI governance work in practice based on the experience and expertise of different countries at various stages of technological and policy development. 

As highlighted during the panel, supervising is a way of re-engineering systems, for example by enhancing them to make room for innovation and new functionalities. This is crucial for the operationalization of legal requirements. It is also particularly relevant in the context of new legal frameworks for the regulation of AI technologies such as the AI Act in the European Union and various normative instruments at the national and international levels that have emerged. While legislation is often seen as a way of re-engineering systems, the focus on implementation is equally vital.

Drawn from the inspiring work of the Dutch Authority for Digital Infrastructure in chairing the European AI Working Group of Competent Authorities, the panel discussion also emphasized the importance of networking and collaboration in addressing cross-border challenges in supervising AI. Panelists unanimously agreed that creating networks is essential to share experiences, overcome obstacles, and make progress in this rapidly evolving area, underlining the significance of collective efforts.

This was followed by a lively discussion between the following panelists:

• David Leslie, Director of Ethics and Responsible Innovation Research, Alan Turing Institute and Professor of Ethics, Technology and Society, Queen Mary University of London;

• Elsa Pilichowski, Director for Public Governance, and Director for Communications (a.i.), OECD;

• JoaquínPérez Catalán, Director of the International Relations Division, Spanish Data Protection Authority, Spain;

• Edson Prestes, Professor at Institute of Informatics of the Federal University of Rio Grande do Sul, and lead expert of the RAM* exercise, Brazil;

• Saida Belouali, Professor of AI ethics, University Mohammed Premier (UMP)and lead expert of RAM exercise, Morocco;

• Julio Alberto PertuzeSalas, Former Under-Secretary for Economics and Small Enterprises and Lead Expert of RAM exercise, Chile;

• Sylvain Faye, Professor, University Cheikh Anta Diop of Dakar (UCAD) and lead expert of RAM exercise, Senegal;

• KristophinaKiito Shilongo, Senior Fellow in Tech Policy, Mozilla Foundation, and lead expert of RAM exercise for Southern and Eastern Africa;

• Samo Zorc, Coordinator of National AI Programme, Ministry of Digital Transformation, Slovenia;

• Hans Hedin, Intelligence Analyst & Intelligence Operations Lead, Swedish Post & Telecom Authority, Sweden.

Following the panel, a session to present the project 'Supervising AI by Competent Authorities' was hosted by the project partners. Participants discussed in depth several supervisory measures, such as standards, audits, and impact assessments to provide their expert insights and thus, input to the project.