IGF 2026 WS #115 AI sovereignty: what for the people?

    Description:

    As artificial intelligence infrastructure expands globally, measures to promote sovereignty in the digital realm often center on states and corporations and on the question of who controls compute, data, and models. This approach risks sidelining the communities whose land, water, energy, and labor sustain this infrastructure. Data centers and the supply chains behind AI (from lithium mining to submarine cables) are frequently sited in the Global Majority or in marginalized territories of the Global Minority, generating environmental and social costs that are unevenly distributed. This session thus asks: what would AI sovereignty look like if rooted in community self-determination, environmental justice, and participatory governance, beyond state or corporate control? We bring together perspectives from academia, civil society, private sector and government to examine how communities affected by AI infrastructure can meaningfully shape decisions about its siting, resource use, capacity-building, and benefit-sharing. The session combines short interventions with a facilitated, interactive discussion designed to surface concrete cases, ranging from data center-related water and energy conflicts to community-led digital infrastructure alternatives. In doing so, the session will stimulate debate on where national sovereignty, corporate interests, and community rights align or conflict, with a focus on territories in the Global Majority. We deliberately seek diversity of geography, stakeholder group, gender, and disciplinary background among speakers and discussants to avoid a one-sided narrative and instead surface genuine tensions and trade-offs. For this purpose, we bring together representatives from multiple continents and stakeholder groups to share their insights and discuss the topic with the IGF community. The session responds to this year's IGF theme by asking who bears responsibility for the social and environmental costs of the "age of intelligence," and how governance frameworks can be reoriented to include those most affected by AI's material footprint.
    Policy Question(s)
    1. How can affected communities shape policies about the technologies and infrastructures they want, as well as their siting, resourcing (energy, water, land), and governance? 2. Where do "AI sovereignty" agendas align with, or conflict against, the self-determination and environmental justice claims of local communities, especially those hosting AI infrastructure? 3. What governance mechanisms can ensure equitable distribution of AI's economic benefits and environmental burdens across the value chain?
    Expected Outcomes
    The session aims to produce a clearer, cross-regional mapping of the tensions between top-down "AI sovereignty" agendas and bottom-up community claims, drawing on concrete cases presented by speakers and audience contributions. Specific outcomes will be the publication of a session report documenting the cases, tensions, and proposed governance approaches discussed, as well as the creation of a network for promoting communities’ participation in the development of policies concerning AI.
    Format
    Roundtable
    Duration (minutes): 90
    Format description: The session will have a fishbowl format. After a short introduction from the moderators, they will invite each panelist to, for up to 5 minutes, answer one of the policy questions according to their experience and expertise. After the first round of inputs, the second phase will last 30 minutes. Anyone from the online or on-site audience can take an empty seat to make a point, challenge a panelist, or bring in a case the panel hasn't covered. After speaking, participants are expected to yield the seat if others are waiting. Moderators will keep the queue moving, periodically restate the policy questions to keep the conversation focused, and actively promote balanced participation in terms of gender, stakeholder group, and geographical origin, keeping the discussion as diverse as possible. In the end, moderators will invite each of the original panelists to name one concrete commitment or next step.
    Hybrid Format: One of the empty chairs of the fishbowl, which will be reserved specifically for participants other than the panelists, will be reserved for a queued online participant, with the online moderator monitoring the chat/hand-raise queue and physically relaying that person into the room via the zoom call on a screen. Moderators will thus alternate: if the last person to take an empty seat was on-site, they will bring in the next queued online participant before taking another onsite one, so the room doesn't default to whoever is physically fastest to move, and vice-versa. Online participants will be frequently invited to actively join the fishbowl, and not just observe it.
    Organizer 1: José Renato Laranjeira de Pereira, University of Bonn's Sustainable AI Lab
    Organizer 2: Felipe da Silva, 🔒Laboratory of Public Policy and Internet - LAPIN
    Organizer 3: Camila Silva, Laboratory of Public Policy and Internet – LAPIN
    Organizer 4: Moraes Thiago, VUB
    Speaker 1: Anita Gurumurthy, Civil Society, Asia-Pacific Group
    Speaker 2: Costa Barbosa Alexandre, Civil Society, Latin American and Caribbean Group (GRULAC)
    Speaker 3: Cecilia Rikap, Technical Community, Western European and Others Group (WEOG)
    Speaker 4: Risper Onyango, Civil Society, Intergovernmental Organization
    Moderator
    José Renato Laranjeira de Pereira, Technical Community, Latin American and Caribbean Group (GRULAC)
    Online Moderator
    Camila Silva, Civil Society, Latin American and Caribbean Group (GRULAC)
    Rapporteur
    Moraes Thiago, Technical Community, Latin American and Caribbean Group (GRULAC)
    What will participants gain from attending this session?