July 8th-11th marked the 2025 AI for Good Global Summit, a United Nations (UN) organized event in Geneva, Switzerland. The event was hosted by the International Telecommunication Union (ITU) in collaboration with several UN agencies. While Atlas applauds international attention to the usage and regulation of AI, the Summit did so in the most ambiguous terms. Starting from the Summit’s YouTube channel highlight videos of each day, speakers picked for the spotlight were high-level figures within the ITU, who merely stated humanity must develop common rules to provide the best possible future for the world. This was continually done, of course, without ever laying out a description of what those common rules or possible future should look like. This leaves the impression that the presence of good intentions will deliver on a good result, which, of course, history is littered with examples to the contrary. Lastly, these videos are characterized more as glamour shots focusing on influential figures and run more like a movie preview than an in-depth rundown of what was discussed at the Summit.
Though the Summit’s Programme provided events with more detail than the highlight videos, a keyword search returned no results for events dealing with “capitalism,” one for “economics,” two for “freedom,” five for “profit,” and several for “governance” and “development.” These are all essential to Atlas’s AI policies, newly enshrined into our Survival Bill. So while we can applaud the ITU and the broader UN for hosting the Summit, we must continue to add pressure to elites to ensure AI development, regulation, and usage results in greater freedom, prosperity, and equity.
Our Survival Bill addresses AI in four primary ways. First, people must be given power over the control and development of AI. We envision this happening through the establishment of a Citizen’s Assembly dedicated to regulating AI for the benefit of all humanity, not the benefit of states and their officials along with capitalists and shareholders, and the drafting of a legally-binding treaty to regulate AI development globally. Second, superintelligent AI development must be temporarily paused until a global governance system for the technology’s regulation can be established along with multilateral verification systems. This would impact only the few companies working on the most advanced AI models, those with the potential to surpass human capabilities. Third, we propose all AI research be ultimately brought under the control of a democratized UN. This would be done through the establishment of the International AI Safety Agency (IAISA) to create rules and enforce decisions, the Global Unit for AI Research and Development (GUARD) to be the sole legal entity to pursue frontier AI research, and an AI Tribunal empowered to settle disputes regarding AI research, development, and ensure compliance. Lastly, the benefits of AI cannot be weighted toward the elites. People must come before profits and to make this a reality we propose an international tax on large AI-based revenues to create and sustain a Universal AI Dividend Fund with the aim of financing global economic, social, and environmental programs.
These solutions are not entirely untested or previously unproposed. Our Italian chapter is working to promote them in their national and sub-national work. As always, if these policies ring true with you we ask that you join us in the fight to make AI both prosperous and accountable to all who will feel its effects.
By Trent Trepanier
Photo: ITU AI for Good - Wikipedia