On July 6, Geneva Hosts the UN's First AI Governance Sit-Down

Starting this Monday, July 6, something a little different kicks off at Palexpo in Geneva, Switzerland. It's called the UN Global Dialogue on AI Governance, and if you haven't heard of it yet, you're not alone. Here's why it matters: this is a platform the UN General Assembly itself set up, and it's the first session where governments, companies, academics, and civil society groups from basically everywhere sit down together to talk about how the world should handle AI. Two days on the calendar, but symbolically it's carrying a lot more weight than that.

Up to now, when people talked about AI regulation, the conversation almost always started with the EU's AI Act or a stack of US executive orders. This time it's different. Instead of one region or one bloc setting the tone, nearly every country on Earth gets a seat at the same table, under the UN's roof. That's part of why some people are already calling this the "Year Zero" of UN-track AI governance. Today we're going to walk through what this Dialogue actually is, who's showing up, what's on the table, and why any of this should register on your radar even if you've never thought about AI policy before.

The Players

Let's start with the two people actually running the show. The co-chairs are Ambassador Egriselda López of El Salvador and Ambassador Rein Tammsaar of Estonia. That pairing is worth sitting with for a second, because neither country is what you'd call an "AI superpower." El Salvador is a developing country in Central America, and Estonia — famous as it is for its digital-government chops — is a small nation of just over a million people. The fact that these two ambassadors are the ones steering the global conversation on AI governance is itself a statement. It signals that this dialogue isn't meant to be owned by the US, China, or wherever the big tech headquarters happen to sit.

Then there's the guest list. Every UN member state government gets an invitation to send a delegation. On top of that, private companies — the ones actually building and shipping AI systems, from Big Tech down to startups — are in the room too. So are academic researchers working on AI safety and ethics, plus civil society groups: human rights organizations, consumer protection advocates, digital rights NGOs, all of them getting a platform to speak. That four-way mix — governments, industry, academia, civil society, all under one roof — is a very UN way of doing things. Nobody gets sole decision-making power; everybody gets a chair at the table.

The institutional backdrop matters too. This Dialogue exists because the UN General Assembly created it as an official track. And in a nice bit of scheduling, the same week in Geneva also hosts the ITU's AI for Good Global Summit and the World Summit on the Information Society (WSIS) Forum. These are different international bodies, each looking at AI from its own angle, and this week they're all converging on the same city. For a few days, Geneva effectively becomes the de facto capital of international AI policy.

One more cast of characters worth naming: the blocs watching from the sidelines. The European Union, which already has its AI Act in force. The United States, which has leaned on executive orders and agency-level actions rather than a single sweeping law. And China, carving out its own regulatory path. None of them are the headline act at this particular Dialogue, but all of them have every incentive to watch closely — because whatever language and priorities emerge here will inevitably bump up against, or complement, their own approaches.

What the Dialogue Is

Let's nail down what this "Global Dialogue" actually is. It's an official platform created by the UN General Assembly — the space where every government, plus the private sector, academia, and civil society, comes together to discuss international cooperation on AI, share best practices and lessons learned, and hold conversations about AI governance that are open, transparent, and inclusive. Those three words — open, transparent, inclusive — are the whole point. This isn't a handful of countries or companies deciding things behind closed doors; it's a declaration that the conversation happens in the open, with everyone watching.

The stated purpose is just as clear. The goal is to make sure AI governance reflects the priorities of every nation — not just the handful that happen to be the most technologically advanced — and to make sure the benefits of AI actually get shared broadly, rather than pooling up around a few countries or companies. That's really the whole reason this Dialogue exists. AI development is concentrated in a small number of places, but its effects ripple out to every country on the planet. This is the UN's attempt to close that gap between who builds AI and who lives with its consequences.

The actual format is a two-day event. Day one opens with a high-level segment — the kind of session where heads of state or ministerial-level representatives give keynote remarks — followed by thematic sessions that are expected to dig into specifics like AI safety, AI and human rights, and how developing countries can actually get access to AI's benefits rather than just its risks. Alongside the main program, there are side events too, giving smaller interest groups room for deeper, more focused conversations on particular issues.

It's also worth mapping this Dialogue against everything else happening in Geneva that same week, since the overlap isn't accidental.

Event Organizer Dates
Global Dialogue on AI Governance UN General Assembly July 6–7
WSIS Forum 2026 World Summit on the Information Society July 6–10
AI for Good Global Summit International Telecommunication Union (ITU) July 7–10

Geneva has always been a hub for international institutions, but this week feels engineered as something closer to an actual AI policy week — with overlapping dates clearly designed so attendees can move between events and keep the conversation going across venues.

What Each Side Wants

Start with developing countries. What they want out of this Dialogue is pretty unambiguous: a real seat at the table where AI development and norm-setting actually happen. Up to now, AI policy conversations have mostly played out as the US, China, and a handful of European countries making decisions that everyone else ends up following. The fact that El Salvador's ambassador is co-chairing this Dialogue reads, to these countries, as a signal that they can actually be participants in shaping the rules rather than just recipients of them. And there's a deeper hope underneath that — that AI's benefits in areas like healthcare, education, and agriculture don't just stay locked up in wealthy countries but genuinely flow outward to them too.

Big Tech and private companies are running a different calculation. What they actually want is predictable, consistent international norms. Right now, regulation varies wildly from country to country and region to region, which means companies have to build out separate compliance systems for every market — expensive and messy. If the UN track manages to establish even a modest set of shared principles, it at least gives these companies a baseline they can point to everywhere in the world. Of course, that also means active lobbying is inevitable, as companies push to keep any emerging rules from tightening too far around their own operations.

Civil society and academia are working a third angle. What they care about most is making sure the risks AI creates — infringement on free expression, algorithmic bias, misuse of surveillance technology, job displacement — get real airtime at the international table, rather than getting quietly sidelined by corporate or state interests. They see themselves as watchdogs in this process, making sure safety and human rights issues don't get pushed to the back of the agenda. For these groups, just having their voices formally recorded in this kind of proceeding carries real weight on its own.

And then there's the UN system itself, which has its own stake in how this goes. This Dialogue is a chance for the institution to prove it's still relevant on the single most important technology issue of the era. If regional blocs keep writing their own AI rules while the UN sits on the sidelines, that raises uncomfortable questions about what the UN is actually for in this space. So there's real institutional pressure for this Dialogue to go well and to evolve into a track with substance — that outcome matters to the UN system well beyond this one event.

Precedents: Wins and Failures

This isn't the first time the world has tried something like this. The comparison that comes up most is the UN's climate framework — the Conference of the Parties, or COP, under the UN climate convention. That process has run since the 1990s, and it has produced genuinely meaningful outcomes, the Paris Agreement being the obvious example. But it's also drawn steady criticism for convening every single year without always translating into binding, enforceable emissions cuts. There's already concern that AI governance could fall into a similar trap — lots of good language exchanged annually, not always matched by real follow-through.

Internet governance offers another useful reference point. The internet has largely run on a "multistakeholder" model rather than centralized control by any one government or institution, embodied for years by forums like the Internet Governance Forum (IGF). That approach gets credit for one clear win: no single government ever managed to seize control of the internet outright. At the same time, it's been criticized for struggling to produce binding norms — the conversation keeps happening, but it rarely lands on enforceable conclusions.

Nuclear non-proliferation offers a different kind of lesson. The Nuclear Non-Proliferation Treaty (NPT) regime, backed by real verification mechanisms like International Atomic Energy Agency (IAEA) inspections, is generally seen as a case where international norms actually gained teeth. It never achieved perfect control over every country either, but it's a useful reference point precisely because it built in mechanisms that actually constrain real-world behavior, not just rhetoric. AI is fundamentally different — there's no physical facility to inspect the way there is with a nuclear site — but the underlying concept of "verifiable commitments" is worth paying attention to as this Dialogue develops.

Taken together, these three precedents frame the real question hanging over this UN AI Dialogue: which shape does it end up taking? A COP-style pattern of talk without teeth? An IGF-style forum that captures a lot of voices but struggles to bind anyone? Or something closer to the NPT model, with actual verification built in? It's genuinely too early to say — the first session hasn't even happened yet — but these historical patterns are exactly the lens worth watching this through.

Competing Approaches

Long before this UN track appeared, regions were already building their own AI regulatory approaches. The furthest along is the European Union's AI Act — a comprehensive law that sorts AI systems into risk tiers and imposes strict obligations on the highest-risk categories. It's legally binding at the regional level, which makes it one of the strongest AI frameworks in existence today. Europe has clearly hoped this model could spread outward and become something close to a global default standard.

The US has taken a different route. Rather than one sweeping law, it's leaned on executive orders and agency-specific actions. The logic is to manage risk without slowing down innovation, but the tradeoff is that policy direction can shift significantly whenever administrations change. And because American Big Tech accounts for such a large share of the global AI industry, US policy choices end up shaping the global AI landscape in practice — regardless of whether those choices carry formal legal weight elsewhere.

China is tracing yet another path — driving AI industrial growth through state direction while simultaneously building out its own distinct rules around content moderation and algorithm regulation. The emphasis differs meaningfully from the human-rights- and privacy-centered approach common in Western regulatory frameworks, and that difference has repeatedly shown up as a point of divergence in international discussions.

Against that backdrop, the real question this UN track raises is whether it's possible to find some minimum common ground that all these different approaches — or countries outside all three — could actually agree on. This Dialogue isn't positioning itself as a replacement for binding regional law like the EU AI Act. It reads more like an attempt to make sure regional frameworks don't collide with each other and that developing countries aren't left out of the conversation entirely. In other words, less a competitor to existing regimes, more a track meant to fill the gaps between them.

So What Changes

Let's break down what actually shifts for each group because of this Dialogue. For government officials and policymakers, there's now one more international reference point to draw on when drafting AI policy. Where the choice used to boil down to "follow the EU model or lean toward the US approach," the principles emerging from this UN track can now serve as an additional axis to consider. That matters especially for developing-country governments that don't have the internal capacity to build AI regulation from scratch — for them, this UN track could realistically become a starting template.

For companies, nothing changes immediately in terms of new binding requirements. This is a dialogue, not a lawmaking session. But over a longer horizon, principles discussed here could plant the seeds for future national regulations or international standards. That's exactly why companies in this space — especially global AI providers operating across many jurisdictions — need to keep an eye on where this conversation is heading. Within a few years, it's entirely plausible that language from this track starts showing up cited directly in national legislation.

For civil society and everyday users, there's no immediate change to feel. Nothing about how you use AI services is going to shift the moment this week's sessions wrap up. But over the longer arc, discussions like this one can gradually build toward international frameworks or baseline human rights standards that offer some recourse when AI causes harm. And for people in less-developed countries who've largely been excluded from AI's upside so far, this Dialogue adds one more legitimate channel for pushing their own governments toward better AI policy.

For researchers and academics, this Dialogue functions as an early signal for how international AI governance conversations will be framed over the next several years — what terminology gets emphasized, which risks get prioritized, and whose voices carry more weight in the room. Watching those patterns play out here gives a reasonably useful preview of where international norms are likely headed. That's why, for people who study this space professionally, the event itself is a data point worth tracking closely, regardless of what concrete outcomes emerge from these two days.

🥄 Three Things You're Probably Wondering

— Is this going to produce any binding agreement? Too early to say. Based on what's public so far, this is designed as a space for dialogue and discussion, not a negotiating table for adopting treaties or legally binding documents. That said, international institutions have a long history of dialogues like this eventually evolving into more formal tracks, so there's real meaning in it as a first step.

— Will anything here actually clash with the EU AI Act or US policy? Hard to say from the published agenda alone. Given that regional frameworks are already in force or well underway, and this UN track is being framed as complementary rather than a replacement, a parallel-track structure seems more likely than a head-on collision. But what actually gets said in the sessions themselves is something to watch closely.

— Will this become a recurring event? This being explicitly labeled the "first session" strongly suggests follow-up sessions were part of the design from the start. But how many sessions are planned, or on what cadence, hasn't been officially confirmed yet — that's something worth watching as more details emerge.

References

Schedule and details are as of announcement and may change.