Global AI Agenda: 17 Nations Chart Course for Government Research

Seventeen countries have jointly endorsed a new government AI research agenda, formalized in the ‘Kyoto Vision for a Golden Age of Science’ declaration. This significant international agreement was signed at the Science and Technology in Society (STS) Forum in Kyoto, outlining a comprehensive framework built around four core pillars: access to scientific data, robust computing infrastructure, unwavering research integrity, and comprehensive technical training.

The declaration primarily concerns government-supported science, focusing on how public institutions will fund research and cultivate the talent required to advance scientific endeavors. A notable aspect of the document is its adoption of the White House’s new terminology, ‘super intelligence’ (SI), to describe advanced AI tools. Ministers believe these tools are poised to dramatically accelerate research tasks and unlock novel forms of scientific inquiry. Michael Kratsios, Director of the White House Office of Science and Technology Policy, emphasized this transformative potential, stating, “Super intelligence applied to science will be the greatest force in history for democratising research. Seizing this moment requires us to reinvigorate our research enterprises and leverage new institutional models for discovery.” The program aims to seamlessly integrate AI capabilities with the resources available to public-sector researchers and the established standards governing scientific work.

A crucial commitment within the Kyoto Vision is the expansion of researchers’ access to SI tools, vast scientific data repositories, advanced computing infrastructure, and essential experimental facilities. These resources are deemed fundamental requirements for conducting research at the ambitious scale envisioned by the declaration. Proposed applications for these enhanced capabilities include sophisticated reasoning across extensive bodies of knowledge, the creation of higher-fidelity models for complex physical and biological systems, and the pioneering of autonomous closed-loop discovery. The White House specifically highlights autonomous experimentation at scale as an anticipated capability of SI when applied to scientific fields. While these are proposed research directions, the declaration refrains from specifying concrete data-access rules or computing allocations, leaving the operational specifics for its access commitments to be defined later. To facilitate this, governments are provided with several funding options, including long-duration awards, rapid grants, and a system of prizes and challenges. Ministers also encouraged experimenting with research organizations that offer varying degrees of autonomy, resources, and direction, alongside supporting dedicated government capabilities in metascience – an empirical field focused on understanding how research is conducted, organized, and funded. This approach integrates the evaluation of research institutions directly into the overall capability-building agenda.

The joint statement also places a strong emphasis on research integrity and technical training, directly linking public trust to scientific rigor. Signatory ministers have pledged to uphold stringent standards for government-supported and government-applied science. These standards encompass reproducibility, transparency, clear communication of error and uncertainty, and unbiased peer review. Significantly, the declaration explicitly recognizes negative and null results as valuable contributions to scientific knowledge. This commitment provides a defined scientific foundation for the trust agenda, valuing findings that do not confirm a hypothesis equally alongside positive results. However, the declaration does not include a separate AI assurance framework nor does it provide evidence that the proposed SI tools currently meet these outlined standards.

Regarding capability building for personnel, the declaration prioritizes early investment in students and early-career researchers based on merit, granting them the necessary support and freedom to pursue ambitious inquiries. Furthermore, it gives particular attention to personnel responsible for operating, maintaining, and improving sophisticated instrumentation, advocating for more opportunities for hands-on experience. Joint doctoral programmes, fellowships, and collaborative research initiatives are suggested as effective pathways for individuals to gain diverse experiences across various scientific environments. These comprehensive commitments collectively address both the human talent and the physical facilities indispensable for conducting cutting-edge research, in addition to ensuring access to necessary software and computing resources.

The seventeen countries that endorsed the ‘Kyoto Vision for a Golden Age of Science’ are: Argentina, Bulgaria, Chile, Cyprus, Germany, Greece, Indonesia, Italy, Japan, Kazakhstan, Republic of Korea, New Zealand, Poland, Singapore, United Arab Emirates, United Kingdom, and United States of America. The White House situates this declaration in alignment with its earlier report, ‘Science: A New Golden Age’, which Michael Kratsios presented to President Trump in July 2026. That foundational report established key directions for US government agencies concerning the conduct and funding of R&D, as well as the measurement of its impact.