IGF 2023 WS #462 Comparing vertical and horizontal approach to AI Governance


Artificial Intelligence (AI) & Emerging Technologies
Chat GPT, Generative AI, and Machine Learning

Organizer 1: Nidhi Singh, 🔒
Organizer 2: Aishwarya Giridhar, 🔒Centre for Communication Governance
Organizer 3: Jhalak Mrignayani Kakkar, 🔒
Organizer 4: Joanne D'Cunha, 🔒Centre for Communication Governance
Organizer 5: Sachin Dhawan, Centre for Communication Governance, National Law University Delhi

Speaker 1: Jhalak Mrignayani Kakkar, Civil Society, Asia-Pacific Group
Speaker 2: Sarah Myers West, Civil Society, Western European and Others Group (WEOG)
Speaker 3: Michael Veale, Civil Society, Western European and Others Group (WEOG)
Speaker 4: Brian Tang, Civil Society, Asia-Pacific Group
Speaker 5: Jessica Fjeld, Civil Society, Western European and Others Group (WEOG)


Nidhi Singh, Civil Society, Asia-Pacific Group

Online Moderator

Aishwarya Giridhar, Civil Society, Asia-Pacific Group


Joanne D'Cunha, Civil Society, Asia-Pacific Group


Panel - 60 Min

Policy Question(s)

A. Given the rapid advancements in the field of AI, how should States approach effective and responsive AI governance mechanisms?
B. How does AI governance differ between Global North and Global Majority countries?
C. What does the intersection of horizontal and vertical approaches to AI look like and how can this be implemented by States?

What will participants gain from attending this session? This panel will provide a platform to convene experts from different jurisdictions who can discuss the trends in AI governance, and chart out future trajectories in this space. Through the session, participants will be able to understand the tradeoffs involved between the two approaches of vertical and horizontal AI regulation, and how both approaches can be used together and separately.
The session will also discuss how these novel approaches can be translated to the Global Majority context, and how they can be implemented there. Attendees from States, civil society, and other stakeholder groups working in the legal and policy space can use this as a starting point while designing their own approaches towards AI governance.
As many countries are currently in the process of adopting their own AI governance mechanisms, we hope participants will be able to then take insights from this session and contextualise it for their own jurisdictions.


As the use of AI systems continues to increase exponentially, the vital question of how to govern AI is being considered by States across the world. The European Union's AI Act will be the first comprehensive legislation of its kind, setting up regulatory frameworks for AI across Europe. This horizontal approach aims to develop broad cross-cutting standards which apply to AI systems across all sectors, and would provide a minimum standard of responsible AI development in the region. Concurrently, States like the US and China are introducing more specific AI regulations covering targeted areas like hiring, algorithmic bias, etc. Such a vertical approach to regulation would produce sector-specific standards to address the specific needs of AI deployment in various use-cases. Consequently it becomes clear that the global AI landscape may involve multiple standards rather than being driven solely by the Brussels Effect.
As different mechanisms for AI develop across the world, it is especially crucial for States in the Global Majority to consider their approach to AI governance, given the disproportionate bias and discrimination that often accompanies the deployment of AI technologies designed and trained in the Global North.
In this session, we will convene experts from across the world to discuss the practical implications of emerging AI governance mechanisms. We will discuss the implications of these approaches in the context of considerations around regulatory capacity, welfare requirements, constitutional and human rights, and economic goals in different jurisdictions. Additionally, we will explore strategies for formulating effective regulatory interventions that can adapt to the rapid pace of AI development and the changes introduced by AI systems. Through this session, we hope to identify the key risks and benefits associated with upcoming approaches to AI governance, which can be used to inform the development of AI regulations across the world.

Expected Outcomes

This session will allow participants to gain insights on the new approaches to AI governance which are currently developing around the world. Participants will be able to understand the underlying distinction between emerging approaches from experts on the panel, assess their risks and benefits and identify ways to adapt this discussion to their local contexts.
CCG will use insights from the session to support the Indian government as it considers its AI governance strategies. As part of the GPAI, CCG will use insights from the session in international policy processes to ensure meaningful representation of Global Majority issues in international norm-formation.
More generally, CCG is a major convenor in the Global Majority space and is a part of both domestic and international processes around AI governance. We will use insights from this session to inform our events and discussions on AI, to allow these conversations to continue beyond IGF.

Hybrid Format: Our session plan is designed to allow for robust engagement and participation, online and offline. To encourage interaction, we will allow significant time for questions and comments from participants once the panelists have made their initial points, and will explore specific areas that participants indicate interest in.
We will provide a variety of ways in which participants can engage - speaking online/ in person, typing questions or comments on the live chat of the Official Online Participation Platform, a public Google document, and a Miro Board. We will encourage in-person participants to also use these tools to engage with those joining online.
Our on-site and online moderators will work to keep track of questions and comments from participants and flag them to the speakers and highlight emerging themes. In addition to active and engaged moderators, we plan to use polls as a tool to encourage participation and information sharing.