The United Nations has apparently decided that the best way to handle the most consequential technology in human history is to put the people selling it in the same room as the politicians who keep getting fooled by it. The UN and its International Telecommunication Union are launching the AI for Good Global Commission, with its first meeting scheduled for July 8 in Geneva. Axios got the exclusive, which means someone at the UN has a very good publicist.
So What Actually Is This Thing
According to Axios, the commission is designed to bring top tech executives and heads of state together to forge global rules for AI. The initiative is framed as a response to the growing splintering of AI regulation around the world, where every country is essentially winging it at slightly different speeds and with wildly different levels of seriousness.
Among those confirmed to be involved: Salesforce CEO Marc Benioff and Rwandan President Paul Kagame. That pairing alone tells you something about the ambition of the project. You have a cloud software billionaire who has spent years positioning himself as the responsible face of big tech, and a head of state who has made Rwanda into something of a showcase for African technological development. It is not the worst starting lineup you could imagine.
The first meeting hits Geneva on July 8, which gives everyone involved approximately one week to pretend they have a plan.
Why Geneva, Why Now
The ITU, the UN body co-convening this commission, is based in Geneva, so the location makes sense. What makes less sense is the timing, if you are being brutally honest about it. The AI industry has already spent three years racing ahead of any conceivable regulatory framework, deploying systems into healthcare, law, hiring, and military applications while governments were still busy asking ChatGPT to write their speeches about regulating ChatGPT.
The EU passed its AI Act. China has its own rules. The United States, as of the moment Axios published this, remains a patchwork of executive orders and strongly worded blog posts. The idea that a UN commission can stitch this together into something coherent is either genuinely visionary or a very well-catered way to produce a non-binding report that gets cited once in a TED Talk.
To be fair, that characterization might be uncharitable. The ITU has done real technical coordination work before. It is the reason your phone works when you travel internationally. Whether that institutional muscle translates to governing systems that can write malware, generate deepfakes, and make parole recommendations is a genuinely open question.
The Credibility Problem Nobody Wants to Talk About
Here is the core tension with any commission that puts AI developers at the same table as regulators: the developers know infinitely more about what their systems can actually do, and they have a direct financial interest in making sure the rules do not slow them down. That is not a conspiracy theory. That is just how incentives work.
Benioff has been one of the more vocal tech CEOs about the need for AI oversight, to his credit. He has called for AI regulation publicly and repeatedly in a way that many of his Silicon Valley peers have not. But Salesforce is also a company that sells AI products and has every reason to prefer a regulatory regime it helped design over one that gets imposed on it later.
This is not a flaw unique to this commission. It is the defining problem of technology regulation in the 21st century. The people who understand the technology well enough to regulate it are almost always the people who built it and profit from it. The UN did not invent that problem. It just invited it to Geneva.
What Would Actually Make This Worth Watching
Let's be clear about what success would look like here, because it is easy to be cynical about multilateral tech governance and also completely correct. A commission that produces a binding framework with real enforcement mechanisms, genuine representation from the Global South, and specific prohibitions on the most dangerous AI applications would be genuinely historic. Not holding breath, but the bar exists.
The Rwanda angle is worth taking seriously. African nations have been largely talked at rather than talked to in previous rounds of AI governance conversations, even as AI systems trained on Western data get deployed in African contexts with often disastrous results. If Kagame's presence means the commission actually grapples with those dynamics, that would be more than symbolic.
Axios notes that the commission is framed as connecting the executives building advanced AI with global politicians. Whether those politicians have the technical literacy, the political will, and the institutional backing to push back on those executives in any meaningful way is the whole ballgame. Everything else is just nice stationery.
The Dingo Take
Here is what we know for certain: AI is being deployed faster than anyone can govern it, the companies doing the deploying have more money and more lobbyists than any regulatory body on earth, and the UN has a distinguished history of producing documents that describe problems with great precision and then mostly leave them in place. That context does not doom this commission. It just means the burden of proof is enormous.
The optimistic read is that someone, somewhere at the UN finally understood that you cannot regulate AI with frameworks built for nuclear weapons or financial derivatives, and that you need a new kind of institution that can move faster and think differently. Putting Benioff and Kagame in the same room is at least a more interesting starting point than another committee of retired ambassadors squinting at a glossary.
The pessimistic read writes itself. A voluntary commission with no enforcement power, operating in a world where the leading AI superpower just spent two years actively dismantling its own oversight infrastructure, is going to produce guidelines that the companies in the room will cite approvingly and then ignore the parts that cost money. We will cover the July 8 meeting. We will try very hard not to say we told you so.
