Meta Science Can Improve Science — But It Must Be Useful to Society}
The emerging field of meta-science aims to improve scientific research itself, emphasizing societal usefulness amid rapid scientific and AI-driven transformations and challenges.


On July 2nd, a scientific initiative was launched in a lecture hall in London, establishing the Meta Science Alliance. Comprising over 25 funding agencies, academic groups, companies, and other organizations, its core mission is to promote the development of meta-science — the use of scientific methods to understand and improve scientific research itself.
The alliance’s formation coincides with a growing community and broader recognition of meta-science. At the Meta Science 2025 conference, over 830 participants from more than 65 countries attended, leading organizers to turn away some applicants.
Amid unprecedented changes in research, meta-science is maturing. Currently, AI is rapidly reshaping research, with budgets tightening and science facing politicization and attacks. Meta-scientists can contribute by pursuing ambitious research to advance science and by applying their expertise to benefit society broadly.
For decades, researchers in scientific and technological institutions have focused on science itself. Since the 2010s, meta-science has flourished, driven by increased awareness of reproducibility and research integrity issues, prompting calls for reform in research evaluation and publication models.
Meta-science now broadly covers peer review, reproducibility, research assessment, impact, open science, and citation analysis. It also involves exploring career paths, funding mechanisms, and addressing inequalities and fairness in science.
In the past year, the field has experienced explosive growth. Tim Errington, Senior Director of Research at the Center for Open Science in Washington, D.C., supports the Meta Science Alliance.
AI in science was a major theme at last week’s conference. Researchers are already using large language models to accelerate everything from funding allocation to evidence synthesis. Meta-science can help document how AI changes science and assist funders and policymakers in strategizing AI’s impact on research.
Last year, the UK launched its first Meta-Science Unit within the Department for Science, Innovation and Technology and UK Research and Innovation. It funds and commissions research to analyze the UK’s complex research environment and find ways to improve it.
Meta-science faces potential misuse, raising the delicate issue of how to communicate the difficulties of reproducibility without undermining scientific credibility. Concealing these issues could worsen the situation, especially as science is under intense scrutiny today.
Scientists must openly discuss systemic flaws and use rigorous methods and data to address them. They should also explore how trust in science is eroding in some regions and how to rebuild it. Improving transparency and communication of scientific principles and results can help counteract efforts to weaken scientific credibility.
Currently, meta-science is at an exciting yet challenging stage. UK Science Minister Patrick Vallance urges scientists to challenge existing systems and respond to public needs, emphasizing that research must serve society — a heavy responsibility.