17 Nations Back Kyoto Vision for AI-Driven Science

The United States and 16 other countries on 4 October 2026 endorsed the "Kyoto Vision for a Golden Age of Science," a joint declaration that commits its signatories to weave AI into how research is done, funded and staffed, the White House announced. The statement was agreed by science and technology ministers in Kyoto, Japan, on the sidelines of the 23rd annual Science and Technology in Society (STS) Forum, and was led by White House Office of Science and Technology Policy director Michael Kratsios. Its central bet is that what the administration calls "super intelligence" will make autonomous, closed-loop experimentation a normal part of science.

The 17 endorsers are Argentina, Bulgaria, Chile, Cyprus, Germany, Greece, Indonesia, Italy, Japan, Kazakhstan, South Korea, New Zealand, Poland, Singapore, the United Arab Emirates, the United Kingdom and the United States. The Straits Times, carrying a Bloomberg report, noted that China was not on the initial list.

What the Kyoto Vision declaration commits to

The full declaration is a single page with three pillars. The first, on improving the institutions of discovery, endorses a deliberate mix of funding tools, including long-duration awards, rapid grants, and prizes and challenges. It encourages governments to experiment with new kinds of research institutions that vary in autonomy and direction, and it welcomes metascience, the empirical study of how science is organized and funded, recommending dedicated metascience capability inside governments. It also pledges reproducible, transparent work that reports error and uncertainty, uses unbiased peer review and treats negative and null results as real contributions.

The second pillar is the AI one. Ministers say AI is already speeding up existing research tasks and opening new forms of inquiry, and they encourage research that uses it to reason across large bodies of literature, build higher-fidelity models of physical and biological systems, and drive "autonomous, closed-loop discovery." They also "endeavor to expand access" to AI tools, scientific data, computing infrastructure and experimental facilities at the scale researchers now need.

The third pillar covers people. Signatories say they will identify and invest early in high-potential students and early-career researchers on merit, expand hands-on training for the technicians who run and improve sophisticated instruments, and encourage joint doctoral programs, fellowships and collaborative research across scientific environments.

How AI is expected to change lab work

Kratsios framed the declaration as a response to a shift that is already under way. "The way we make scientific discoveries has changed, and the Kyoto Vision is a recognition of that shift and a call to action," he said in the White House release, adding: "Super intelligence applied to science will be the greatest force in history for democratizing research."

In remarks reported by The Straits Times, Kratsios sketched what that looks like in practice: automated cloud laboratories operating at industrial scale, where robotic systems run hundreds of experiments in parallel and feed results straight back into analysis and hypothesis refinement. He also said machines can synthesise thousands of papers and compress years of work into "tens of hours" for teams of digital agents assisting human researchers.

That vision overlaps with debates PromptCrates has followed closely. Researchers have warned that automating AI research and development itself could trigger an intelligence explosion that outruns oversight, while applied tools such as Microsoft's Quine system, which helped scientists prioritize pancreatic cancer compounds, show the narrower, near-term version of AI-assisted discovery that the declaration is trying to scale.

Safety concerns voiced on the same stage

The Kyoto meeting was not purely celebratory. At a side event the same day, SoftBank Group chief Masayoshi Son, speaking alongside Kratsios, said even he is worried about safety as machine capabilities climb quickly, according to the Bloomberg report in The Straits Times. Son said countries need to build trust and work together to control the threat from powerful models, and warned that super intelligence in the wrong hands could become "super dangerous." The report notes that SoftBank has directed nearly US$65 billion into OpenAI.

Son also predicted that AI will move into a phase of helping robots and other machines act in the physical world starting in 2027, and repeated his forecast that AI and learning machines will account for at least 20% of global GDP by 2040, The Straits Times reported.

The safety note sits awkwardly with the declaration itself, which says almost nothing about risk. The text mentions public trust and scientific integrity, but it does not address how autonomous labs should be supervised, how dual-use findings in biology or chemistry would be handled, or who is accountable when an AI system proposes a dangerous experiment.

What it means for researchers and funders

For working scientists, the most concrete promise is access. If governments follow through on expanding researcher access to AI tools, data, compute and experimental facilities, smaller labs and universities could gain capabilities that today sit mostly with large technology companies and well-funded national labs. The emphasis on technicians and instrument skills also recognizes that autonomous labs still need people who can keep the hardware running.

The limits are equally clear. The declaration is non-binding: its verbs are "endeavor," "encourage" and "welcome," and neither the White House release nor the text attaches any money, timelines or shared infrastructure to the commitments. The White House says the Kyoto Vision builds on Science: A New Golden Age, a report Kratsios delivered to Trump in July 2026, which set new directions for how US agencies conduct, fund and measure research.

The geopolitics deserve attention too. Indonesia, Singapore, Japan and South Korea all signed, giving the declaration a strong Asia-Pacific footprint, while China's absence underlines that AI-for-science is becoming another arena of bloc-building. PromptCrates has covered the parallel track in the US and China super intelligence dialogue.

What to watch next is whether endorsing countries translate the text into budget lines: dedicated metascience units, shared compute for academic researchers, or funding calls for autonomous laboratories. Without those, the Kyoto Vision risks being remembered as a well-phrased statement rather than the start of a new model for science.

• The White House: US leads international coalition to endorse Kyoto Vision for a Golden Age of Science
• The White House: Kyoto Vision for a Golden Age of Science (PDF)
• The Straits Times (Bloomberg): SoftBank's Masayoshi Son has rare cautionary note on AI safety
• PromptCrates: AI researchers on intelligence explosion oversight
• PromptCrates: Microsoft's Quine system and pancreatic cancer research
• PromptCrates: US and China super intelligence dialogue