All presentations are online and the recordings from the meeting will be made available in mid-June 2025.
The COAR Conference took place in Tokyo, Japan on May 12-15, 2025 and was attended by approximately 140 people representing at least 23 different countries. The meeting was jointly organized by COAR, JPCOAR, and the National Institute of Informatics (NII) and provided an opportunity to learn more about the current open science landscape in Japan as well as exchange with others about national and regional trends, challenges, and strategies for repositories and repository networks.
Conference participants heard that in Japan, an open access policy has recently come into effect that requires researchers to deposit their published articles into a Japanese repository. Japan has a very robust repository network of about 700 repositories, many of which are hosted by the NII national infrastructure and JPCOAR, a large community of repository managers, has taken on an active role in helping to raise awareness of the policy on their campuses.
The infrastructure in Japan for open science is very robust because of significant investments made over the last decade, with well-developed data and institutional repositories. However, like in many countries, open science requires a cultural change. Japanese researchers are not yet fully aware of the requirements of the policy and there is significant work to be done to socialize the policy within the research community. There are also challenges related to integrating Japanese research outputs into the international corpus of scientific knowledge because of the distinct aspects of the Japanese language and non-latin characters.
There were also presentations from many other regions and countries around the world including Spain, Canada, India, Czechia, Latin America, United States, South Korea, Africa, Turkey, Australia, United Kingdom, France, and the European Union. Presenters were asked to identify the top two recent developments related to open science in their country and the impact this was having on repositories. While each country / region has their own unique context, three main trends emerged out of these presentations:
Repositories have become the preferred compliance mechanisms for open access and open science policies. Many current policies require the deposit of published articles into a repository, unlike a decade ago, when policies were typically agnostic about how researchers can make their articles openly available and there was no stated preference between publishing directly in OA or depositing a copy of a manuscript into a repository. This has increased the perceived value of repositories world wide, but has also put a strain on some who may not be immediately prepared to handle an increasing volume of papers, navigate opaque and confusing rights retention environments, and explain policy requirements to researchers.
There are increasingly significant actions being taken to reform research assessment systems. Research assessment reform is needed in order to incentivize researchers to practice open science, rather than simply publish in high-impact journals. Europe, in particular, is making significant inroads with transforming research evaluation frameworks to include open science practices as important criteria. In Spain, for example, open science practices represent 10% of the general assessment criteria.
Countries are expanding their investments in locally-hosted and governed infrastructures. Instead of focusing on transformative agreements because of their prohibitive costs, countries are moving investments towards open infrastructures. Open access repositories, which evolved as institutional services are developing shared solutions to reduce costs and redundancies. There are also important efforts to improve the interoperability across open science infrastructures, which typically have been disconnected and not aligned in their activities. This allows countries and regions (e.g. EOSC) to connect publication and data repositories, diamond journal platforms, indexing and aggregation services, and research assessment systems creating a more integrated, coherent open science landscape.
Several other important topics were also discussed at length during the conference and are summarized here:
The emergence of AI, with its huge potential and but also many challenges
Discussions around AI are ubiquitous these days and this was also a topic that was addressed at the COAR Conference. Several scenarios were also discussed in terms of the tremendous potential for AI and Large Language Models to add value to large aggregations of open access literature from repositories. For example, could AI be used to process multilingual content in a way that it would be accessible to speakers of any language or could it be used to provide plain language summaries of research outcomes. There are also now many AI tools that are being developed to assist individual repositories with improving the quality of their repository services such as helping with the metadata curation, detecting relationships between related entities (papers, authors, institutions and funders), and enabling new forms of user engagement.
On the flip side, repositories are also now frequent targets by AI bots who want to scrape / download their content. These bots are sufficiently aggressive that they can cause service disruptions and outages in repositories. The community discussed the results of a COAR survey that found that open repositories are being significantly impacted by AI bots and other crawlers. This issue poses a challenge for repositories, who are sometimes responding by blocking bots which also interferes with their mission to provide FAIR and open access to their collections. A number of solutions were proposed at the meeting, such as whitelists and bot filtering and in June, COAR will launch a AI-bots task group to consider various solutions and develop recommendations for the repository community.
Despite these issues, there is a growing consensus in the community that open access repositories have a competitive advantage over the publishers and other closed providers, when it comes to developing scalable AI solutions based on our collective open access content.
The value of high quality metadata for building good and reliable network services
Metadata is the foundation for almost all of the networked services that are built on repository content. Therefore the value of these services is directly related to the quality of the metadata on which they depend. Meeting presentations highlighted a range of services including access and discovery, research assessment and tracking, and the reuse of content in particular using AI and automated tools to interpret and process data without human intervention—enabling large-scale meta-analyses, predictive modeling, and real-time insights. Even with more automated ways of improving metadata quality through various tools and integration with other systems such as PIDs and aggregators, this is still not sufficient for quality service development and repositories must invest more in their metadata curation activities.
The growing acceptability of the “publish, review, curate” publishing model
The publish, review, curate (PRC) model is a method of publishing research that is built on open access manuscripts deposited into repositories and with external peer review performed by an overlay journal or other open review service. COAR has been advancing P,R,C as an exciting alternative to traditional diamond publishing that introduces innovation across the publishing workflow including publishing of a variety of types of research outputs and alternative peer review processes. In addition, it is a low cost option with a smaller technical burden for the research community. The COAR Notify protocol was developed to enable seamless interactions across PRC initiatives and is now being widely adopted by repositories and other infrastructures.
COAR would like to sincerely thank the National Institute of Informatics (NII) and JPCOAR for providing such excellent hosting and organization of the meeting. The interesting debates, discussions and information sharing will contribute significianly to the themes and topics in the next COAR Strategy that will be developed in the summer / fall of 2025.
