NEW YORK — Bill Gates is warning technology companies and world leaders that they must act now to ensure artificial intelligence combats social inequities rather than deepening disparities — and his foundation is pledging US$1 billion to improve access. The Microsoft co-founder made the appeal in the Gates Foundation’s annual Goalkeepers report, which tracks progress on the development goals set by the United Nations to improve quality of life by 2030. The donation announced Tuesday will be spread over the next two years for AI-focused efforts to train models on local languages, improve health outcomes, upgrade educational tools and inform small farmers’ practices around the world. And despite the United States’ international aid cuts, Gates Foundation CEO Mark Suzman said he would continue pressing U.S. Congress and U.S. President Donald Trump’s administration to lead the way on foreign assistance. “We haven’t given up on the U.S.,” Suzman told The Associated Press in an interview. “We will acknowledge it’s been a challenging period with the U.S. government on this. But we believe there is still a case to be made.” The long-prepared report arrives as executives at leading AI companies urge others in the industry to slow its fast-moving development. An ex-Anthropic researcher’s resignation over concerns that developers are not acting responsibly has revived the debate over whether increasingly powerful AI models could escape human control. Gates shared similar concerns in a post last month on his website. AI will either be “the greatest equalizer” or the “worst source of injustice,” he wrote. He acknowledged that the technology will empower people to carry out fraud, disinformation, surveillance, cyberattacks and bioterrorism. Annual road map published at time of heightened anxiety over AI The Gates Foundation, one of the world’s largest global health funders, is focused less on funding safeguards and more so on democratizing what Gates considers to be the technology’s benefits. Gates pledged in last month’s blog post that “all of our work” will “take full advantage of AI,” which is one of several focus areas benefiting from this year’s record budget. But the market will design the best AI tools first for those who can afford them, Gates suggested. “That is how inequity compounds: the technology that makes life easier, increases productivity, and sometimes even saves lives arrives first for people who already have the most,” he wrote in the Goalkeepers report. The Gates Foundation views AI as “a potentially transformative tool” for a “world that is in a little bit of crisis,” according to Suzman. The foundation projected 2025 to be the first year this century with an increase in child deaths in last year’s report. The Trump administration’s disruptions to a global AIDS relief program have impeded efforts to prevent and treat the disease in dozens of countries. Meanwhile, the gap between rich and poor nations is growing even wider, according to a recent U.N. report. For AI to “help accelerate progress,” Suzman said, it must perform well in all local languages — not just English and, to some extent, Mandarin. The foundation announced Tuesday about US$100 million to help build data sets in the languages spoken by underserved communities and reflect their local contexts. More data won’t guarantee more equitable tools, according to Jonathan Shaffer, an assistant professor of sociology at the University of Vermont. Shaffer, whose research focuses on how global health governance is shaped by science and knowledge, noted that health infrastructure is extraordinarily weak in many parts of the world. The most marginalized populations would likely remain excluded from the data fed into health-oriented AI models, he suggested, which “may even amplify the kinds of inequalities” the foundation aims to reduce. Working with OpenAI, Anthropic and other tech giants on health and agriculture The Gates Foundation is turning to some of the biggest AI companies. Google.org and Microsoft’s AI for Good Lab helped launch an open funding call for projects strengthening the frameworks needed for thousands of underrepresented languages across Africa. OpenAI committed US$50 million alongside the foundation to train health workers in Rwandan clinics as part of a pilot Suzman said will soon include other African countries. Anthropic is working with the foundation to accelerate vaccine developments and improve its chatbot’s data set of local crops, among other efforts. Tuesday’s announcement promised to bring about US$400 million more for education, such as a tool that helps teachers rework lesson plans around the concepts least understood by their classes; some US$400 million for healthcare, such as a tool that double-checks clinicians’ reports for overlooked symptoms; and roughly US$100 million for agriculture, such as an app that offers real-time advice to farmers on weather patterns, pest control and fertilizer usage. Suzman acknowledged these totals are a “tiny proportion” of the billions of dollars spent by tech giants on their commercial products. He confirmed the foundation is having additional conversations with the major AI companies about how their technical resources can contribute to these goals and where their leaders should invest the vast individual fortunes they’re expected to make. Gates also noted a “potential bias” in his August blog post: he still has financial ties to the technology industry, where he built a fortune estimated at more than US$100 billion, the bulk of which he has pledged to the Gates Foundation. He also works alongside Microsoft and other AI companies as the Gates Foundation’s chair. Still, he emphasizes that his views are not profit-motivated. Shaffer, the University of Vermont sociologist, questioned how equitable these efforts can be when its deployment is largely directed by private foundations and corporations instead of “democratically organized, deliberative bodies.” “They still mean a kind of particularly privatized form of access,” he said. Discussions at next week’s United Nations gathering The Goalkeepers report comes ahead of the United Nations General Assembly’s high-level meetings next week in New York. U.N. Secretary-General António Guterres has attempted to include developing nations in conversations about AI governance. Guterres, whose term leading the international body will end this year, has said the technology “cannot be governed by a handful of countries or companies.” But U.N. processes “move considerably less rapidly than AI does,” according to the International Crisis Group’s Richard Gowan, a program director and U.N. watcher. The U.N. also faces questions about its relevancy as it closes in on the selection of Guterres’ successor. “It isn’t clear that the next secretary-general will keep up this focus,” Gowan told reporters on Monday. ___ James Pollard, The Associated Press
