IGF 2024 Open Forum #38 Harnessing AI innovation while respecting privacy rights

Adopted in 2019 as the first intergovernmental standard on AI, the the OECD Recommendation on Artificial Intelligence (AI) set a global reference for human-centred AI development and deployment, establishing principles to promote AI innovation that respects democratic values and human rights, including privacy rights. The Recommendation was revised in 2024 to ensure it reflects and addresses important technology and policy developments, notably with respect to general purpose and generative AI, and heightened concerns around privacy, safety, information integrity, and environmental sustainability. The Recommendation complements OECD standards in other areas such as privacy, notably the OECD Privacy Guidelines, whose principles form the foundation of privacy laws globally.

Rapid advancements in AI capabilities, including the rise generative AI, have raised questions around the privacy implications of AI associated with the training and deployment of AI systems, and how to protect the rights and interests of all parties affected by these developments. This Open Forum will report on the key revisions to the OECD Recommendation on AI and the rationale for the changes, highlighting key privacy implications. While privacy was already considered in the 2019 OECD AI Principles, the updated text calls out privacy explicitly in the heading of one of the Principles (1.2) to align with OECD policy priorities in the context of generative AI.

The Open Forum will also seek to foster a discussion on the intersection of AI, data, and privacy, leveraging the work of the OECD.AI Expert Group on AI, Data and Privacy. The discussion will focus on questions such as:

• What was the rationale for updating the OECD AI Principles and what are the main changes? In particular, what are the revisions pertaining to privacy?
• With the rise of advanced AI systems, notably generative AI, how does the evolving demand for data affect privacy concerns?
• In balancing AI advancement with safeguarding privacy, what policy measures can be implemented to fully unlock the potential of AI while ensuring privacy protection?
• What is the interplay between privacy legislation and AI-specific regulations? How do they complement each other and what is the interaction between their respective safeguards?

Agenda

1. Setting the scene (5 minutes) - The moderator will describe the purpose of the open forum and introduce the panellists.
2. Panel discussion (30 minutes) – The moderator will invite the panellists (four) to discuss the questions and provide examples/best practices (e.g. OECD taxonomy of data collection methods for AI).
3. Questions and answers (20 minutes) - The moderator will open the floor for questions and answers between the audience and panellists.
4. Wrap-up (5 minutes) – The moderator will summarise the main points and wrap up the discussion.

For the interactive Q&A sessions, the online moderator will make sure that online participants are as able to take part in the discussion as onsite participants. The online moderator will oversee the chat and ensure equal participation by equally splitting the question time for onsite and online participants. The OECD will be able to advertise the event in the social media pages of the OECD.AI Policy Observatory, on the communities of the Digital Policy Committee, of the Working Party on Artificial Intelligence Governance, as well as the Working Party on Data Governance and Privacy. In addition, the OECD will advertise the event through its network of OECD.AI experts.