IGF 2023 WS #561 Learning from IG Multistakeholderism to govern AI


Artificial Intelligence (AI) & Emerging Technologies
Chat GPT, Generative AI, and Machine Learning
Future & Sustainable Work in the World of Generative AI

Organizer 1: Herman Ramos, 🔒Inaglobe
Organizer 2: Samik Kharel, 🔒Independent
Organizer 3: Harimino Lalatiana Rakotondrainibe, 🔒
Organizer 4: Mohamed Salah, 🔒Canar

Speaker 1: Morales Angulo Cristina de los Angeles, Civil Society, Latin American and Caribbean Group (GRULAC)
Speaker 2: AVIT NDAYIZIGA, Civil Society, African Group
Speaker 3: Samik Kharel, Civil Society, Asia-Pacific Group
Speaker 4: Mohamed Salah, Technical Community, African Group


Herman Ramos, Technical Community, African Group

Online Moderator

Harimino Lalatiana Rakotondrainibe, Civil Society, African Group


Harimino Lalatiana Rakotondrainibe, Civil Society, African Group


Break-out Group Discussions - 90 Min

Policy Question(s)

A. What can we learn from the current process of the Internet Governance Model and apply it to AI regulation?
B. Will it be legit for AI to function in the same pattern as Multistakeholder Model?
C. How the current functions of Internet governance and stability be enhanced or challenged by the rapid developments of AI?

What will participants gain from attending this session? This session offers critiques understand of how multi-stakeholder models are applied both in theory and practice for discussion involving AI, and how it can allow to reframe the discussion and explore innovative approaches in tools, frameworks, and concepts.
Participants will gain an understanding of factors determining whether (and in what form) multistakeholderism is practiced in the development of responses to guide the development and use of trustworthy AI for the wider good and how is applied for other particular issue-area, over time and the appropriate criteria for matching governance modalities to particular governance functions.
Therefore, through this session, participants will gain insights into the long-term normative vision for the AI developments that are taking shape and how to establish a multistakeholder approach to allow the participation and cross-border collaboration of different stakeholders in shaping an enabling policy environment by putting in place regulation and guidance that promote trustworthy AI.


Rapid technological development has increased the complexity and interconnectivity of both networked systems and information and communications technologies (ICTs). AI has become one of the most researched subjects in recent years special due to its potential to offer new solutions in different areas. Nevertheless, AI can also amplify and entrench current social, economic, and geopolitical problems.
Not all countries around the globe are equally prepared for the impact of AI, and some countries have been struggling to adapt to the rapid pace of change AI brings. For developing nations, the AI revolution can be a game changer after the missed opportunity with the industrial revolution. The isolated work between governments and organizations can originate in the development of ambiguous and different plans that can affect not only the implementation of AI but also harm the user. No single country or stakeholder has all the answers to the challenges present by AI and collaboration between all stakeholders of internet governance is fundamental, because progress is only possible when different organizations work together.
The multistakeholder model can offer diversity and pledges individual, local, and regional voices. This model allows the change from states, and treaties between states, towards non-state actors and the inclusion of different stakeholders for different regions for better articulation, and execution of regulation, laws, and norms that serves all intervenient.
The development and deployment of AI demand new and effective policy responses. However, the multistakeholder approach is a time-consuming process, while AI is so fast-moving. This session explores how this engaging process can be a suitable approach for AI regulation considering its interactive undertakings.
The concept of multi-stakeholderism is further disaggregated into a typology that distinguishes several forms based on the varieties of actors involved and the nature of authority relations between them.

Expected Outcomes

As an outcome is expected a document that will be published contain a complete analysis of how the Multistakeholder model can help address both narrow issues implicated by the institutions that govern the technical AI and the broader issues related to accountability, ethics, etc.
This session will result in a guide that will be used for training different stakeholders on how the multistakeholder model can help ensure consistency across jurisdictions, because of the wide agreement regarding AI principles and values.
Also, the report produced will contain analyses of how this model can help that foster innovation and trust in AI by promoting the responsible stewardship of trustworthy AI while ensuring respect for human rights and democratic values.
The outcomes and document produced by this session will contain clear guidance from multilateral bodies and governments on when and where regulation on AI tools is needed based on collaboration with different stakeholders.

Hybrid Format: The session will be hybrid to ensure that all attendees can participate on equal footing. The organizers are prepared to ensure the best possible experience for online and on-site attendees. The session is designed so that it can start with an introduction, in the sequence, the speakers will start the discussion, process, and engagement with the participants. Then we will have a considered number of minutes of exchange conversation between the speakers and attendees.
Then will proceed with the second part of the discussion which can include break rooms in order to develop a plan that meets the goal of the session and align with the topic and will finish with an interaction with all participants.
The moderators onsite and online will manage the session and will gather contributions from the attendees. Moderators will encourage attendees to effectively allow them to be part of the real-time conversation.