The 2027 Swiss AI Summit is still taking shape, and that is an opportunity. On 6 May, the MAP-AI initiative, led by the Centre for Communication Governance (CCG) at National Law University Delhi and the Global Network Initiative, brought together a wide range of diverse stakeholders from five regions to help make the most of this opportunity. The online workshop, titled “Collective Stewardship: Multistakeholder Roles in AI Safety Summits,” was co-hosted with Switzerland’s Federal Office of Communications (OFCOM), who joined to share where planning for the 2027 Summit currently stands and, just as importantly, to listen to the regional learnings and perspectives from past summits and global processes.
This event was originally planned as a private gathering at RightsCon in Zambia. While GNI and CCG were disappointed not to have the opportunity to hold it in person, we were pleased that over 65 participants were able to join virtually. OFCOM noted that details about format, date, and scope of the Summit remain open and explained that they are actively working through how to make sure the Summit is inclusive, well-organized, and impactful. No official date has been fixed, but the Swiss are looking to host the event in Q2 of 2027, with a precise date to be announced in due course. They explained that a public curtain-raiser for the Swiss Summit will be held at AI for Good Global Summit in Geneva in early July 2026, where they plan to share more details. OFCOM also noted that the United Arab Emirates and Japan have both expressed interest in hosting future summits in the series.
On the broader ambitions for the Summit, OFCOM was clear that the goal is to continue strengthening global AI governance and to help establish continuity for the outcomes and work of this summit series, while being careful not to duplicate existing forums such as the UN Global Dialogue on AI Governance and AI for Good Global Summit. They are thinking carefully about how to institutionalise the summit series in a way that adds value rather than creating overlap. Ensuring that voices of people and companies from the Global South are heard and involved was flagged as a particular priority, with OFCOM noting that lessons from digital governance processes can help inform how to make the Summit as multistakeholder as possible. Geneva’s unique ecosystem was also highlighted as a distinctive asset, with OFCOM expressing a clear intention to engage the wide range of international organisations based there, including the Red Cross, International Telecommunications Union, International Labour Organisation, World Intellectual Property organisation, and the International Meteorological Organization.
Participants in the MAP-AI workshop came from civil society organisations, research institutions, policy networks, the private sector, and governments across regions. Breakout groups were organized by region (Africa, Latin America, Asia, Europe, and North America) and worked through four shared prompts that together offered a practical framework for thinking about the priorities, framework, processes, and outcomes for the 2027 Summit.
Each group was asked to reflect on: (1) what the Summit’s agenda should prioritise, (2) how multistakeholder participation and inclusion can be structured and funded, (3) what accountability mechanisms would make commitments meaningful, and (4) which stakeholders and processes the Swiss should be engaging with in the lead-up to 2027. Across all five regional groups, the responses were specific, grounded in experience, and notably consistent.
Priorities: anchor the Summit in human rights
The priority that came up first, and kept coming back across every region, was human rights. Not as a vague aspiration, but as a concrete and binding framework for understanding state and private sector responsibilities and anchoring key commitments and deliverables. Participants welcomed the opportunity the Swiss Summit presents to center the international human rights framework within AI governance. The UN Guiding Principles on Business and Human Rights, EU AI Act and other emerging normative standards were seen as useful reference points to anchor understandings of private sector responsibilities, governments duties, and how Summit commitments are framed.
Participants also saw real potential in broadening the Summit’s view of what constitutes an AI risk. In addition to longer-term frontier risks, which have been the focus of many AI governance processes, near-term and concrete harms must be prioritized. As participants noted, these include: biometric surveillance, data extracted from communities without consent, and the energy and water costs of data centers. Examples of priorities that emerged from the regional breakout groups included African participants making a strong case for treating African data as a resource that must be protected, pointing to the large-scale extraction of data used to train AI systems without community consent, and calling for gender equality to be mainstreamed as a cross-cutting principle rather than addressed as an afterthought. They also pushed for the Summit to go beyond declarations on democratising AI, asking how data and compute can be made genuinely accessible to lower-income economies in practice.
Structural questions about who benefits from AI and who absorbs its costs, including through rare earth extraction and data labour, were also considered ripe for serious discussion at a summit of this scale. From a Business and Human Rights perspective, participants from Latin America pressed for the Summit to include Human Rights Due Diligence as a concrete expectation of private sector actors, not just a reference point. From Asia, participants raised the importance of building continuity across summits, asking whether progress on those commitments would be tracked and carried forward in Geneva.
Participants from North America also added an often-overlooked dimension: the role of investors, including the World Bank, IFC, and development funders, in shaping how AI is developed and deployed globally, and the importance of ensuring that this community is meaningfully engaged in Summit discussions.The participants also noted that the 2026 India AI Impact Summit established important foundations around democratising AI resources and multilingual AI, and they are keen to see the Swiss Summit carry those threads forward and deepen them.
Participation: reach the people who need to be there
The India AI Impact Summit in February 2026 featured participation from Global Majority organisations in a way that earlier summits had not. The question participants were keen to explore is how the Swiss Summit can build on that progress and go further.
The most consistent ask was practical: fund participation. Visas get denied, travel is expensive, and the organisations doing the most relevant work on AI’s real-world impacts rarely have the budgets to get to cities like Geneva. Dedicated funding for civil society travel, alongside active support with visa processes, would make an immediate and tangible difference to who is in the room. Participants from the Global Majority were clear that this is a prerequisite, not an add-on. Latin American participants echoed this directly, noting that enabling Global South participants to physically get to Geneva is foundational to any meaningful participation strategy.
Beyond logistics, participants were enthusiastic about formalising civil society’s role in agenda-setting. Several pointed to the IGF’s intersessional model, the WSIS+20 model and the NetMundial+10 São Paulo Principles as frameworks worth drawing on. A process where civil society organisations help shape what gets discussed, not just who attends, would mark a real step forward. The North America breakout group raised a pointed question that resonated across regions: can Switzerland formalise a mechanism for civil society participation in agenda-setting itself, and be transparent about how those inputs feed into decisions? Transparency about how inputs feed into decisions matters too: knowing that a regional consultation actually influenced an outcome is both motivating and accountable.
The Africa breakout group welcomed the Swiss government’s early signals about possibilities for skills and capacity transfers between African and Swiss universities, and suggested that research emerging from such collaborations could be showcased in the lead-up to the Summit, connecting the event to the research and development aspirations embedded in many national AI strategies across the region.
On format, there was genuine appetite for working groups open to multistakeholder input, hands-on workshops, and co-creation sessions over conventional panels. These formats create space for smaller organisations and those outside established networks to contribute as peers. A multilingual approach, with materials and key sessions available beyond English, would extend that reach further. Several participants also called for dedicated space and funding for civil society side events, and for mechanisms that allow for meaningful remote participation across time zones.
Performance on commitments: build accountability from the start
Participants saw a real opportunity for the Swiss Summit to raise the bar on how commitments are made and tracked. The basic building blocks are clear: a public register of commitments with named owners and timelines, mid-cycle reviews, and some form of independent monitoring. Several models were discussed, including the UPR’s traffic-light tracking approach, the GNI accountability framework, and the Extractive Industries Transparency Initiative, with broad agreement that a lightweight but transparent mechanism is both achievable and overdue. The North America breakout group also raised a longer-term vision of a mechanism inspired by GNI, with annual reviews, named members, and structured accountability procedures, while acknowledging that a lighter-weight commitment tracker might be a more achievable starting point.
The ask was not for a new bureaucracy, but for commitments that are designed to be accountable from the outset, with specific language, named stakeholders, and a clear plan for what happens between summits. The Asia breakout group suggested that accountability mechanisms could be built into session proposals themselves, so that every commitment made during the Summit comes with a stated owner and review process attached. The Latin America breakout group pushed for commitments to move beyond general principles toward enforceable language across four specific areas: human rights safeguards, digital sovereignty, environmental accountability, and supply chain responsibility, with strong language accompanied by implementable actions.
One idea that generated real energy was connecting Summit commitments explicitly to other processes such as the UN Global Dialogue on AI Governance, national AI strategies, and human rights reporting cycles, so they become part of an ongoing accountability ecosystem rather than stand-alone pledges. The Latin America regional breakout group also named “summit proliferation” directly, noting that the growing number of overlapping AI governance processes is itself a governance problem, and that the Swiss Summit has an opportunity to bring greater consistency by building on the IGF’s intersessional model to sustain stakeholder engagement between events. The North America regional breakout group added a useful framing: how can existing transparency and accountability mechanisms, many of which do not yet adequately account for AI, be built upon and layered into rather than duplicated? The Swiss Summit is well placed to help build that connective tissue, given Geneva’s unique concentration of international institutions. Rather than adding to the proliferation of overlapping governance efforts, it could help organise and link them.
Parties and processes: show up before the Summit
On coordination, the spirit of the discussion was generous. The Swiss have a real opportunity to build trust and relationships well before the Summit begins, by engaging in the spaces where relevant communities already gather. AfricaIGF, LAC-IGF, CIPESA’s FIFAFRICA, Paradigm Initiative’s DRIF, the Digital Rights Asia Pacific Assembly, and the Open Government Partnership summit in the UK are all places where the conversations that should feed into 2027 are already happening. The UN Global Dialogue on AI Governance is another key space where participants identified opportunities for collaborative engagement. Participants in the Asia regional breakout group flagged this space in particular, noting that the Swiss need to find a way to engage constructively without drawing energy away from that process, alongside OECD discussions.
The EU regional breakout group pointed to Brussels as a critical hub for engagement on the AI Act and sovereign AI debates, alongside the Council of Europe’s Steering Committee for New and Emerging Digital Technologies work on implementing HUDERIA, the European Humanitarian Forum, and the Global Gathering in Portugal. The North America regional breakout group suggested hosting something during UNGA high-level week to reach a broader international civil society audience.
There was also enthusiasm for reaching communities that do not always see themselves as part of the AI governance conversation: health workers, labour markets, environmental advocates, and humanitarian assistance providers. AI is already shaping their fields. Participants also pointed to the value of highlighting emerging regulatory models from diverse contexts, including Brazil, Kenya, and South Korea, that are navigating these questions in instructive ways. Bringing those perspectives into the Summit would make it more grounded and its outcomes more credible.
What comes next
The curtain-raiser in Geneva in July is the next key moment on the road to the Swiss AI Summit. Meanwhile, MAP-AI will continue to create structured opportunities for civil society to input into the 2027 Summit planning process and continue to make the case for meaningful multistakeholder participation in AI governance. This includes developing clearer expectations for what civil society engagement in processes like the 2027 Summit should look like in practice, grounded in the experiences and priorities of communities across the Global Majority. We are also committed to deepening and growing this community, creating more opportunities for organisations across regions to connect, collaborate, and show up to these processes with a strong, unified, and more impactful voice.
We are grateful to the Swiss Government for their collaboration and genuine openness to engagement, as well as to all those who participated in this event. We look forward to working with the broader MAP-AI community and the Swiss Government to make the 2027 Swiss AI Summit an impactful example of multistakeholder participation in AI governance.