The Future of Archives of Science: An Interview with Polina Ilieva

In the new open-access volume Archives of Science: Challenges and Opportunities in the 21st Century, editors Polina Ilieva and Venkat Srinivasan bring together archivists and scholars from around the world to confront a central question: How can archives keep pace with the evolving practices of science?

The volume ranges from case studies on the use of AI in archival workflows to ethical dilemmas in medical archives, and from data visualization projects that make hidden figures visible to a reappraisal of foundational theories for documenting scientific records.

I was delighted to have the chance to interview Ilieva about this new book. In what follows, she reflects on what inspired this global collaboration, why she believes this is a critical time for renewed investment in documenting science, and how new tools and frameworks might shape the next generation of archival practice.

Trevor Owens: The book’s introduction emphasizes both “challenges and opportunities.” What do you see as the most pressing challenge facing science archives today—and what opportunities most excite you?

Polina Ilieva: Thank you for this opportunity, Trevor. Before we dive into the questions, I want to take a moment to express my heartfelt gratitude to my co-editor, Venkat Srinivasan, and the general editor of the University of California Health Humanities Press, Brian Dolan, whose invaluable support made this publication possible. I also want to thank all the authors from across continents and countries who generously shared their insights and experiences. This book stands as a testament to the openness and inclusivity of the archivists, librarians, and museum professionals who work tirelessly to preserve and ensure access to the archives of science for all.

The challenges and opportunities are deeply intertwined. Contemporary science is evolving at a breathtaking pace, it’s more than ever team-based, collaborative, and increasingly open, with preprints, shared datasets and code, electronic lab notebooks, and even podcasts and blogs shaping the research landscape. Yet archival practice is struggling to keep up, especially when it comes to capturing not just the final products of research but the full, dynamic process by which science is made.

The most pressing challenge is scale and complexity: how to document and link the entire research lifecycle: data, software, protocols, products, and discourse across distributed, often ephemeral platforms, while ensuring authenticity, provenance, privacy, and long-term sustainability.

While many tools have emerged over the past two decades (as authors in our volume discuss, from AI-assisted appraisal and processing to new methods for archiving robotics and other complex instruments), they are often developed within individual institutions. Even when released on open-source platforms, they’re not universally adopted, and they can become unsustainable when the core development teams leave or funding shifts. This leads to fragmentation and tool churn, further widening the gap between rapidly evolving science and the archival infrastructure meant to preserve it.

The opportunity that excites me most is co-developing archival tools and practices with the scientific community and repurposing institutional research systems for archiving. We should build an interoperable infrastructure that connects data repositories, publications, lab records, and traditional archival collections that were digitized and became computationally actionable by using persistent identifiers, shared metadata, and linked open data, so researchers and the public can explore an integrated “research universe” rather than siloed environments. Equally important is establishing cross-institutional governance, maintenance plans, and sustainable funding models to ensure these tools endure beyond individual teams and institutions. This connected, durable ecosystem is where archives can truly add value.

TO: In your co-authored chapter, you revisit the 1985 guide Appraising the Records of Modern Science and Technology . What motivated you to return to that foundational work, and what did you discover through the gap analysis and interviews with scientists for this project?

PI: The 1985 guide, Appraising the Records of Modern Science and Technology, was a groundbreaking framework for understanding how scientific records should be preserved, but science itself has evolved dramatically in the 40 years since its publication. Revisiting this foundational work allowed us to explore the transformations in how science is conducted, who participates in it, and how its results are communicated.
One significant shift has been the rise of collaborative science. While the original guide acknowledged collaboration, the scale and complexity of scientific teamwork today is unprecedented. Another major change is the movement toward openness in science. Many scientists, both in basic and clinical research, are advocating for greater transparency, open access, and a reimagining of scholarly communications. However, the dominance of traditional publishing models, such as subscription-based journals, continues to present challenges. Open access publishing, while transformative, remains an expensive endeavor, and navigating these costs requires systemic change.

We’ve also seen shifts in how scientists communicate their findings. While publishing in prestigious journals remains the gold standard, many researchers are increasingly using social media platforms to share methods, amplify their work, and engage with broader audiences, including the general public. Social media, for example, YouTube, has become a valuable tool for making science more accessible and promoting reproducibility.

Through our gap analysis and interviews with scientists, we found that, despite the adoption of new communication channels like social media, major journals and conferences continue to be regarded as the primary venues for disseminating research. Social media, on the other hand, is often used more informally to share methodologies, connect with peers, and increase visibility.
Our team, representing three diverse institutions, is currently preparing an article that will explore all the findings of this research.

TO: Several chapters highlight efforts to engage contemporary scientists—sometimes through documentation strategies, other times through collaboration on access and interpretation. What approaches or goals do you see as most promising for archivists and librarians seeking to work more directly with the scientific community?

PI: One of the key goals for archivists is raising awareness, i.e. ensuring that scientists within our institutions and beyond understand the role the archives play in preserving and providing access to these valuable materials. This can be achieved through organizing in-person and virtual events, exhibits, and sharing stories online that showcase the research value of our collections.

For example, at my institution , we hold one of the most extensive collections documenting the AIDS epidemic . Over the past several years, we’ve hosted young researchers as part of their summer fellowship programs and recently collaborated with a lab led by a principal investigator studying mechanisms that promote effective and durable CD8+ T cell immunity to viral pathogens (e.g., HIV, SARS-CoV-2). Their lab was specifically interested in exploring earlier research on the HIV virus. This type of engagement emphasizes the importance of introducing scientists, particularly early-career investigators, to archives and their value for contemporary research. This is especially critical for materials generated during periods when data management practices and requirements for data sharing were not yet standardized.

While archivists serve as custodians of these materials, collaboration with subject specialists is essential for ensuring their relevance and accessibility. For instance, Dr. Diane Havlir has recently provided a historical overview of key figures and findings in HIV/AIDS research at UCSF, which now serves as a “collection development” guide for our archives and strengthens our partnership with scientific experts.
In addition to these efforts, the completion of mass-digitization projects has allowed us to make archival materials computationally actionable, following an “archives as data” approach. With the advent of AI, these digitized collections that were once hidden or locked in paper format are finding new audiences and applications. This demonstrates the transformative potential of technology in unlocking historical data for contemporary scientific and interdisciplinary research.

However, engagement often starts with overcoming barriers. For instance, when approaching scientists with renowned careers, we frequently hear them ask, “Would anyone be interested in what I did and how I did it?” These conversations highlight the need to start outreach early, before scientists reach the peak of their careers. Including information about archives in records management guidelines, even briefly, could help foster awareness and engagement.

The blog I’m currently interviewing with plays a crucial role in bridging the gap between scientists and archivists. By bringing these communities together, it fosters dialogue, collaboration, and a shared understanding of the immense value archives hold for contemporary research. It highlights how archivists and scientists can work hand-in-hand to preserve history and make it actionable for future discoveries.
Finally, I believe that archives and libraries are uniquely positioned to host digital humanities or health humanities centers. These centers can serve as hubs to connect historians, anthropologists, engineers, basic scientists, and other researchers with tools and approaches for utilizing computationally actionable historical data. By facilitating collaboration across disciplines, we can ensure that our collections are not only preserved but actively used to advance knowledge and research.

TO: Born-digital records emerge as a major theme across the volume. Given that most contemporary science is now fully digital, how does the book help readers think about preserving born-digital scientific records? And how has your own perspective on this evolved through the work?

PI: Currently, I would estimate that 80% of our collections are hybrid, with a growing portion being fully digital. A notable example of a born-digital collection is the COVID Tracking Project , a citizen science initiative that was run during the first year of the pandemic collecting and publishing data required to understand the COVID-19 outbreak. To process this collection, we collaborated closely with the team that originally managed that project to develop archiving tools specifically tailored for its unique requirements of preserving social media, code repositories, oral histories and other data. While we’ve made considerable progress in processing and even preserving such collections, the biggest challenge lies in providing seamless access to them.

Access to these digital records is facilitated through multiple platforms, including Archive -It for websites, Dryad for data , Calisphere , HathiTrust , and others. However, the challenge is ensuring this access is streamlined and user-friendly. To address this, with the support from the Sloan Foundation our team is developing a secure virtual reading room to enable access to the COVID Tracking Project materials, leveraging innovative technology developed within our institution.
Digital preservation is often overlooked or lacks a programmatic approach. At our institution, we are addressing this gap by utilizing in-house technology for the preservation of born-digital data.

TO: The volume also explores new technologies, from interactive data visualization to AI-assisted processing. What examples from the book do you find particularly exciting, and how has your view of these tools’ potential (and limitations) changed through the editing process?

PI: Both examples you highlighted are particularly compelling to me, especially in the context of mass digitization of historical records and their transformation into computationally actionable data. For instance, in my institution, UCSF, the recently digitized collections of five women scientists who played pivotal roles in establishing the field of child development offer immense potential for mapping and network analyses to understand how the field evolved. Additionally, I see significant opportunities in leveraging AI for metadata creation and processing in both digitized and born-digital collections. However, ethical challenges persist, particularly regarding metadata that is processed by technology but lacks contextual depth.
I’m especially intrigued by the experimentation described by the authors from Caltech in their chapter on AI-enhanced appraisal and processing of hybrid collections [Chapter 6]. Their exploration of integrating machine learning into archival workflows shows promise in automating traditionally time-intensive tasks, such as identifying duplicate records, extracting metadata, creating finding aids, and analyzing thematic content. At the same time, our colleagues at the National Center for Biological Science (NCBS) in Bangalore, India, have raised critical ethical concerns regarding the risks of blanket digital access, which can inadvertently enable AI-driven unauthorized reuse of sensitive information [Chapter 10].

Several authors also emphasized the importance of data visualization. For example, the American Philosophical Society (APS), adopted a relatively low-tech approach to dataset creation, relying on researchers to read every letter and review every document within archival collections [Chapter 11]. I found their suggestion to enhance available data at APS by incorporating Linked Open Data, such as Wikidata, particularly interesting. This approach not only enriches the archival data but also creates opportunities for cross-archival collaboration, where archives can contribute to a shared pool of linked data.

Another area of interest is the future potential for archiving software critical to the preservation of robotics, which would expand the Carnegie Mellon team project [Chapter 9]. Emerging technologies, including AI and advancements in digital preservation, are opening unprecedented avenues to amplify access, enhance discovery, and reimagine the role of archives in serving diverse communities across disciplines and geographies.

However, technology alone cannot drive progress. As we experiment with these tools, ethical considerations must take center stage. Archives are custodians of cultures, scientific heritage, and history, and deploying new technologies without thoughtful reflection risks distorting the truth, eroding trust, or prioritizing convenience over ethical stewardship. The challenge lies not merely in innovating but in innovating responsibly, ensuring that principles like transparency, accountability, and inclusivity guide every technological advance.

TO: Looking ahead, what impact do you hope Archives of Science will have on the field? What conversations or collaborations do you hope it inspires in the next decade?

PI: Looking ahead, I hope Archives of Science becomes a catalyst for earlier, closer collaboration between archivists, scientists, and our colleagues in scholarly communications and research data management. I want it to inspire us to innovate and to experiment—whether that’s new ways of documenting research, new metadata practices, or new approaches to preservation and access.

Hearing from peers across institutions and countries helps us spot patterns we might otherwise miss: persistent gaps in the history of science, blind spots in network analysis, and the many “hidden figures,” especially women and historically marginalized communities, whose contributions have been overlooked. Our role is to surface those stories, so no one remains invisible, and to do so with a global lens that includes science from all countries.

New technologies, especially AI, carry real promise but also ecological and ethical challenges. I hope Archives of Science attests that libraries and archives are laboratories, a space to test, critique, and responsibly leverage these new technological tools that balance innovation with responsible stewardship.

We began these conversations in 2024 at the 4th Workshop on Scientific Archives , and we’ll carry them forward at the next meeting in 2026 at the University of Coimbra in Portugal. My hope is that over the next decade, these continued discussions about the archives of science will result in long-lasting collaborations, shared standards, and open access that will make the scientific record more complete, more connected, and more inclusive. I would like to invite all readers of the Ex Libris Universum blog to explore this open-access volume and add to these discussions.