Global Law for AI Governance could protect all humanity by November 20, 2025
By Craig Warren Smith Chairman, Digital Divide Institute
Even after a full year of worldwide efforts to regulate Artificial Intelligence, it remains unregulated. Moreover, no plan has been proposed to regulate AI globally. But here is a plan, what I call AI Moonshot, drawing upon lessons learned in the 30-year-long movement to close the digital divide. The aim of the plan is to contain Artificial Intelligence (AI) before it controls humans. I believe we have only 18 months to get it done—after which the $100 trillion USD AI Gold Rush sweeping the West will overwhelm and distort any plans for AI containment. This introductory article is the first of a five-part series presenting the AI Moonshot plan in detail. I’m not saying the plan is likely to be realized, but there is a slim chance it might be. It shoots for the moon to clarify aspiration and stimulate discussion and debate.
Here are the plan’s characteristics:
It would quell geopolitical discord between the US and China and within Russia-Ukraine and the Middle East—the three locations now threatening to converge and break out into World War III.
It may stop global warming by redirecting the cause from the ineffective approach of the United Nations and establish global warming as the leading social cause of the new era of Generative AI.
It would be affordable by leveraging, augmenting, and redistributing the trillions of dollars that AI will predictably bring to commercial markets.
It would reduce inequality between and within nations, which is already accelerating as AI governance continues to elude the world.
It would accelerate the shift of the Middle East away from an oil economy to a regional digital economy, in sync with plans of Saudi Arabia and the United Arab Emirates.
It would save tens of millions of jobs that may otherwise be forever lost and elicit critical thinking, reskilling, and upskilling at scale.
It could foster collaborations between the world’s top AI research labs by producing incentives that foster collaborations between the world’s #1 innovators (in America’s Silicon Valley) and the #2 cluster of innovators (in China and its diaspora), though currently both sides pursue a self-defeating winner-takes-all strategy.
It would shift the centrical force of the AI era towards the Global South so that it is no longer dominated by a few affluent nations of the Global North. It would look beyond saturated consumer markets of the North to open pent-up demand for products and services to establish a four billion-strong global middle class. It would give starring roles to the Champions of the Global South in three countries—India, Brazil, and South Africa, and in Indonesia.
Here is a peek at how all the above would be achieved by the end of 2025:
The AI Moonshot plan will tap into the platform of the G2I (which is the Group of 20 nations plus the 55 nations of The African Union included as one member for the first time). The G21 now represents 90% of global GDP and is the closest thing that exists to a functioning world government—with enough votes in the UN Security Council and UN General Assembly to produce an international AI law with breadth and staying power. By historical coincidence, four developing nations have been and will continue to be G21 hosts in a year-to-year series from 2021 to 2025, ending in the G21 November meeting in Johannesburg. They are: Indonesia, India, Brazil, and South Africa, with a combined population of more than 2 billion spanning five continents—enough to activate economies of scale for an AI ecosystem. In fact, they could set the agenda for USA, China, and the EU (as one member of G21).
Within the G21, a little-known subgroup called the Working Group on Digital Economy incorporates AI into its framework and is committed to establishing continuity on AI policy from one year to the next. This could emerge as the platform for the AI Moonshot. The plan for the moonshot has these five steps towards its realization:
The first money for building this moonshot would be $500 million or so, coming from a coalition of international foundations—Ford, Rockefeller, Carnegie, et al. The purpose is to conduct a neutral best-practice assessment of data distribution, digital economies, and AI in the four target nations.
Create a secretariat to achieve a rapid series of seminars in all six nations and the nations that surround them in six continents, leading to a Digital Economy Working Group of G21 coalition.
Work through the foreign ministries of six nations to conduct seminars and craft detailed reports tied to the Working Group on Digital Economy of G21—culminating in a seminar on November 16, 2025, just prior to the G21 leaders' meeting on November 21, when Biden, Xi, and other heads of state could endorse the moonshot.
This then allows the coalition of nations to conduct difficult deliberations during the next 12 months in pursuit of these themes:
AI risk containment in the short, medium, and long term
Global warming and AI
Education upskilling and reskilling
Financial scenarios for AI moonshot—where the money will come from
International law.
Leading to the announcement of the new international law when the South African President convenes the leaders on G21 in late November 2025.
Here is a summary of the next four articles in the series over four months:
Article 2: A Middle Way in AI: Not as regulated as China; Not as deregulated as the USA. Here is an optimal scenario for what global AI governance would look like.
Article 3: Financing the global AI ecosystem. Sam Altman wants 7 trillion for chips. Cathie Wood of ARK Investment thinks $100 trillion will flow to big AI. Here's how the World Bank would want $100 trillion to upgrade the digital economy so that all would benefit.
Article 4: Education and Jobs: Here's how remaking education and jobs will lead to sustained growth in emerging markets.
Article 5: Rethinking International Law.
Comments