Huge tech firms are deploying compute clusters with hundreds of thousands of GPUs to coach and run AI fashions. However throughout the complete continent of Africa — encompassing 54 international locations and greater than 1.5 billion folks — fewer than 1,000 GPUs can be found for researchers and builders to coach fashions on local-language datasets.
That disparity illustrates what the United Nations’ high digital envoy calls an “immense focus of tech energy and wealth” in a couple of zip codes — not simply international locations or areas, however confined areas, primarily within the U.S., the place the businesses shaping AI are primarily based.
He didn’t title names, however the level hit near dwelling for the Seattle viewers: 98109 for Amazon, 98052 for Microsoft.
Delivering the keynote deal with through Zoom at Seattle College’s 2026 Ethics and Tech convention on Friday, Below-Secretary Amandeep Singh Gill known as 2026 “particularly seminal” for AI governance, because the expertise shifts from mannequin capabilities and infrastructure funding to techniques that carry out real-world duties autonomously.
Gill pointed to the international response to Anthropic’s Mythos AI mannequin — which the corporate restricted from broad public launch over cybersecurity considerations — for example of why AI governance requires a complete, worldwide method.
Listed here are extra of the important thing messages from his speak.
AI may grow to be a “systemic threat.” Gill stated the expertise is a “comparatively minor threat” now however warned it may quickly bypass cybersecurity defenses, speed up armed battle, and erode public belief by means of deepfakes and misinformation. “Once we can’t inform the distinction between what’s true and unfaithful, what’s actuality or imaginary, then we lose this shared sense of an understanding of info,” he stated.
Armed battle may worsen. Gill warned that AI dangers “reducing the brink of battle, complicated accountability below worldwide humanitarian regulation, and setting us off on escalation ladders that we can’t management.”
AI’s vitality calls for are threatening local weather objectives. The vitality required for giant language fashions, agentic techniques, and inference is already threatening nationwide net-zero targets, Gill warned. Information middle emissions, water consumption for cooling, {hardware} turnover, and mineral extraction prices are compounding — and falling disproportionately on low-income international locations.
AI is each “a possible answer and a stressor” for the atmosphere. It may optimize renewable vitality grids and speed up progress in fusion and batteries, however the short-term prices are mounting. Gill stated the UN is inspecting how to make sure fairness and simply transitions “over these time horizons.”
The UN is constructing a scientific panel for AI modeled after the Intergovernmental Panel on Local weather Change. Chaired by journalist and Nobel Peace Prize winner Maria Ressa and Turing Award-winning AI researcher Yoshua Bengio, the 40-member panel is intentionally composed of solely two members every from China and the U.S., with the remaining 36 from different international locations, together with seven from Africa, to make sure extra international locations are heard. Its first report is predicted in July 2026.
The UN is placing AI governance conversations below one roof. Conversations about AI beforehand occurred in separate our bodies with slim mandates. Now they’re being introduced onto what Gill known as a “horizontal platform” the place policymakers from all 193 international locations can study from one another and develop widespread approaches.
Gill known as AI governance a “sovereign choice.” The UN received’t inform international locations how one can regulate AI, however governance frameworks imply little if nations lack the capability to take part. Gill known as for help of community-driven AI tasks that spend money on native analysis and innovation ecosystems, permitting folks to make use of these instruments to resolve their very own issues.
He acknowledged the UN is working with restricted sources in opposition to an infinite problem, however stated the choice is leaving AI’s trajectory to market forces and geopolitical competitors.
The objective, he stated, is a world the place AI empowers democracies and societies, and creates alternatives not only for “a couple of billionaires and trillionaires” however for everybody.
