Metascience Meetup in Tokyo 2026
On March 14–15, 2026, a two-day event Metascience Meetup was held at Miraikan in Tokyo. Below is my quick take.
The program featured three keynote lectures and five sessions, with somewhere between twenty and thirty speakers in total. Professor James Wilsdon gave one of the keynotes via video. He was kind enough to mention my “Strands of Metascience” diagram.
The other keynotes were by Rei Nouchi, a philosopher of science who spoke on epistemic pluralism and the conceptual resources that philosophy can bring to the metascience movement, and Takahiro Miura, a Science of Science researcher who described his field as a “science toolsmith” that provides diagnostic instruments for the health of the research system.
The five sessions covered international trends in metascience, collaboration with Science of Science, AI for Science (the session I organized), lightning talks from researchers in fields ranging from philosophy to public policy to cell biology, and a critical examination of open science that included perspectives on when not sharing data is the right thing to do.
Over a hundred people came from academia, government, and industry. I have been introducing the metascience movement since around 2023, and it was genuinely moving to see this many people gather for an event under the name of “metascience.”
Why do we need to talk about metascience at all? I think the short answer is that science is A) a social process, and B) it changes. Science produces knowledge collectively (A), which means people can’t just make up their own rules. There has to be agreement on how science is done: what counts as valid knowledge, what kinds of knowledge are worth pursuing, how individual contributions are evaluated, who provides resources and how. And because science keeps changing (B), that agreement can’t be settled once and for all. It has to be renegotiated as technological and social conditions shift.
Traditional metascience, philosophy of science for instance, has pursued these questions out of intellectual interest in their own right. But the contemporary metascience movement asks something broader: that people involved in science and research step back from simply doing it and think about how it ought to be done collectively. Not as a new academic discipline, but as a cross-cutting effort. Professor Wilsdon calls it a “big tent,” Nouchi a “tag” you can attach alongside your own expert domain .
Get Ryuichi Maruyama’s stories in your inbox
Join Medium for free to get updates from this writer.
What I found myself thinking about over the two days was the scale of unit at which this consensus gets formed. Different fields have different circumstances, and several speakers pointed out that different countries may hold quite different ideals of what a scientist should be. But what struck me was that the communities sharing a common way of doing science may end up even smaller than nations or fields.
Take one example. There are now researchers who have AI agents conduct research and write papers while they themselves serve as the person responsible for the work. Some are moving toward that way of working. Others hold firmly to the principle of writing their own papers. Once that kind of divergence takes hold, the meaning of a research paper, and how you read one, inevitably starts to differ from person to person. In the AI for Science session, Akiko Aizawa described how researchers are shifting from being writers to something more like editorial directors overseeing AI-generated drafts, while Makoto Kureha raised the philosophical question of whether, if AI-driven prediction works without explanation, the value we place on scientific understanding itself might erode. These are not distant possibilities. They are happening now, and they are pulling people’s assumptions about what science is in different directions.
As social expectations for science change, as AI reshapes the tools of research, and as the number of researchers worldwide keeps growing, values and objective of science may well fragment. If they do, evaluating research by a single standard becomes harder and harder. To my eyes, science today is subject to exactly that kind of centrifugal force.
Many people from the science and technology policy community, from MEXT and JST, attended the meetup, which I found enormously encouraging. At the same time, I could feel how difficult the situation must be for them. A MEXT official spoke candidly about the challenge of policymaking when the object of policy itself is becoming harder to pin down, and about the ministry’s efforts to move past an assumption of infallibility and make room for experimentation. But if science and scholarship fragment at the level of values, the difficulty runs deeper still: in central government policymaking, you may not even be able to define the problem you are trying to solve.
Kazuhiro Hayashi closed the event by saying that metascience should aim to “understand by creating the practice of science.” I think this follows naturally from the situation I have been describing. If you can no longer hold on to the idea of one science, or even a small number of them, then rather than trying to catalog all the emerging variants, it may be faster to simply build the kind of science you believe in. The question is at what scale this should happen. Science is a social process, so one person can’t do it alone. But maybe fifty people or so can. At that scale, you propose the way of doing science you think is right, gather funding, and try it out. This resonated with what Ryosuke Shibato of Academist Inc. described in his talk: a research ecosystem in which research driven by researchers’ own visions springs up in a distributed way. Rather than trying to impose a grand consensus on a landscape where new modes of science keep appearing, the task for metascience and for policy is to encourage that diversification while cultivating the conditions for the better approaches to grow.
The above is my very abstract takeaway from the event. We should be part of the effort to build a new mode of science while trying to bridge between the modes. And we should try to have a space for dialogue between different scientific and non-scientific communities. That is what I think my role will be going forward as a “metascience communicator.”
