Artificial intelligence is rapidly moving beyond commercial applications and into the heart of scientific research. In a significant international development, 17 countries have endorsed a new agenda aimed at using advanced AI systems to accelerate government-supported science and research.
The agreement, known as the “Kyoto Vision for a Golden Age of Science,” was endorsed at the Science and Technology in Society Forum in Kyoto. It focuses on improving access to scientific data, computing resources, research infrastructure and technical talent.
The participating governments see advanced AI as a potential catalyst for a new era of scientific discovery. Rather than limiting AI to administrative or productivity tasks, the agenda explores its use for complex scientific problems.
AI could help researchers analyze enormous collections of scientific knowledge, create sophisticated models of biological and physical systems, and potentially support autonomous experimentation and discovery.
However, these remain proposed directions rather than confirmed large-scale implementations. The declaration does not yet establish specific rules for data access or detailed allocations of computing resources.
Better Access to Data and Computing
One of the major priorities is giving researchers greater access to the infrastructure required to work with advanced AI.
This includes:
- Scientific datasets and research databases
- High-performance computing infrastructure
- Experimental facilities
- Advanced AI tools
- New research funding mechanisms
The countries also support different approaches to research funding, including long-term grants, rapid funding programs, prizes and scientific challenges.
The goal is to create an environment where researchers can experiment with AI while having access to the computing power and scientific infrastructure needed for ambitious projects.
Research Integrity Remains a Priority
While AI can dramatically accelerate research, the participating governments also highlighted the importance of maintaining scientific standards.
The declaration emphasizes reproducibility, transparency, peer review, communication of uncertainty and research integrity.
Importantly, the agenda also recognizes negative and null scientific results as valuable contributions. This is significant because AI-generated research will need to be evaluated using rigorous scientific standards rather than simply being accepted because it is produced quickly.
Investing in the Next Generation of Researchers
The initiative is not focused solely on technology. Developing scientific talent is another major component.
The countries support greater opportunities for students and early-career researchers, along with doctoral programs, fellowships and international research collaborations.
There is also an emphasis on people who operate and maintain sophisticated scientific equipment. As laboratories become increasingly automated and AI-powered, skilled researchers and technical specialists will remain essential.
Which Countries Signed the Vision?
The 17 countries supporting the Kyoto Vision are:
Argentina, Bulgaria, Chile, Cyprus, Germany, Greece, Indonesia, Italy, Japan, Kazakhstan, South Korea, New Zealand, Poland, Singapore, United Arab Emirates, United Kingdom and United States.
The initiative reflects a broader international trend toward cooperation on AI for science. Earlier international AI discussions have similarly emphasized research investment, workforce development, international collaboration and responsible AI development.
What This Could Mean for the Future
The biggest potential impact could be the emergence of AI-powered scientific laboratories, where AI systems assist researchers in designing experiments, analyzing results and identifying new research directions.
If governments successfully combine AI models with high-quality scientific data, advanced computing and physical laboratory infrastructure, research cycles could become significantly faster.
At the same time, the success of this vision will depend on practical issues such as access to computing, data governance, scientific validation, cybersecurity and the ability to train enough researchers to work effectively with AI.
