Open Science vs. Research Security

As governments tighten research policies, datasets disappear from public repositories, and artificial intelligence consumes unprecedented volumes of scientific literature, a fundamental question is emerging: how open should science be? In this episode of SIAcast, Jonny Coates and co-host Katharina Miller bring together experts in open science, privacy, licensing, research integrity, and research security to explore whether openness and security are truly opposing forces — or whether both depend on something deeper: trust. Through discussions on Creative Commons licensing, health data governance, AI training, research sovereignty, and scientific collaboration, the guests challenge the idea that science must choose between openness and protection. Instead, they argue for a more nuanced approach — one that recognizes the need for secure, trustworthy systems that enable sharing while safeguarding people, knowledge, and institutions. From privacy-by-design principles to the future of AI and international collaboration, this conversation offers a thoughtful roadmap for navigating one of the most important debates in modern research. Key Takeaways Open science and research security are not inherently opposites. Effective innovation requires both openness and appropriate safeguards, including cybersecurity, data integrity protections, and trusted collaboration networks. Trust is the foundation of both privacy and openness. Without trust, patients, researchers, companies, and institutions become less willing to share data and collaborate, ultimately weakening research outcomes. Creative Commons licenses provide flexibility, not unlimited openness. They allow creators to determine how their work can be reused while preserving agency over sharing decisions. Anonymization alone is no longer sufficient. AI-driven inference and data linkage create new privacy risks, making context, governance, and layered safeguards increasingly important. Overly restrictive research security policies can backfire. Limiting international collaboration may isolate researchers from cutting-edge developments and reduce national scientific capacity. AI is reshaping what “open” means. The large-scale use of openly available content for AI training has raised new questions about attribution, reciprocity, and the sustainability of the commons. Biographies Co-host Katharina Miller is a lawyer, ethicist and foresight strategist who advises companies, public institutions, universities and international organizations on AI governance and responsible innovation through her firm, Miller International Knowledge. She also teaches at IE University and serves as an ethics expert for the European Commission. Guests María Bardají Cruz is an independent global privacy and AI governance advisor focused on the responsible use of data in research and innovation. She examines how privacy, trust and robust governance frameworks can support the responsible advancement of research and the use of data for societal benefit. Richard Gold is a legal scholar and innovation policy expert whose research focuses on knowledge sharing, intellectual property, and the role of openness in scientific progress. He advocates for collaborative approaches that strengthen both innovation and research sovereignty. Monica Granados works with Creative Commons on issues related to open knowledge, licensing, and the future of sharing in the age of AI. Her work focuses on creator agency, attribution, and sustaining the commons through effective governance. Leslie McIntosh is a research integrity and scientometrics specialist with a background in biomedical data systems and research governance. Her work examines trust, knowledge security, and the institutional responsibilities required to maintain trustworthy science. Chapter Timestamps 00:00 Introduction: Why openness, security, and trust matter now 01:00 Richard Gold: Why open science and research security need each other 04:23 Monica Granados: How Creative Commons licenses balance openness and control 07:08 María Bardají Cruz: Privacy, trust, and health data governance 09:20 AI, anonymization, and the limits of privacy protection 12:30 Where should the limits of openness be for health data? 15:07 Leslie McIntosh: Rethinking research security as knowledge security 17:00 Misconceptions and resistance to research security policies 22:02 National security, scientific leadership, and international collaboration 24:17 The risks of overly restrictive research security measures 26:51 Protecting datasets with Creative Commons licenses 28:06 AI, data sharing, and the future of openness 30:51 How AI challenges traditional ideas of open licensing 34:28 Creative Commons Signals and attribution in AI 37:36 Who is responsible for research security? 40:19 Practical steps for universities balancing openness and protection 41:45 Looking ahead: openness, privacy, security, and trust in 10 years 43:35 Final reflections: trust as the foundation of science and innovation Music: Upbeat Corporate by JP Bianchini. Produced by Rippling Ideas for the Science Integrity Alliance.