Strengthening Secure and Ethical Data Visitation in FAIRlyz: A Testing Ground for the EOSC-Future/RDA AIDV-WG’s Recommendations
Adoption Factsheet
Outputs Adopted: Guidance for Informed Consent in the Context of Artificial Intelligence and Data Visitation, AI Bill of Rights Recommendation, Guidance for Ethics Committees Reviewing AI and Data Visitation
Related RDA Groups: Artificial Intelligence and Data Visitation (AIDV) WG
Adopting Organisation: FAIRlyz
The RDA Output adopted
The Artificial Intelligence and Data Visitation (AIDV) WG contributes to building the ethical, legal, social, and technical frameworks for Artificial Intelligence while examining the potential of data visitation to bridge challenges for the open sharing and re-use of data in the framework of Open Science. It ams to bring together expertise across disciplines and regions to ensure ameliorate the use of AI in research and innovation across technologies and sectors to address the grand challenges of society while promoting an understanding of how data visitation can contribute to improved data sharing.
The Guidance for informed consent in AI and DV examines the role of informed consent in AI and DV, addressing fundamental challenges to current informed consent frameworks and practices. The aim is to provide guidance for researchers and data controllers across disciplines regarding informed consent in AI and DV.
The AI Bill of Rights promotes fundamental human rights and advances trust in AI and federated systems for Open Science.
The Guidance for Ethics Committees Reviewing AI and DV outlines basic ethical principles that will assist research ethics committees to understand the questions, methods, and procedures for reviewing Artificial Intelligence (AI), including use of Data Visitation (DV) in research with humans.
This Adoption Story was created with the support of the RDA TIGER project, as part of the RDA TIGER Cascade Grants programme.
What was the challenge that you addressed?
This RDA Adoption Story presents the TIGER‑supported DV4RDA project, which implemented outputs from the EOSC‑Future/RDA Artificial Intelligence and Data Visitation Working Group. These standards enable trustworthy AI and secure data‑visitation workflows, supporting compliance and privacy‑preserving research infrastructures. They address the growing need in biomedical research for robust governance and automated data‑quality assessment as AI increasingly works on sensitive, distributed datasets. FAIRlyz, a data‑quality platform by Lifetime Omics built around secure data visitation, applies these principles by integrating legal, ethical, and governance requirements into its central registry and QC tool, which runs directly within users’ computing environments.
The adoption process
The project adopted AIDV‑WG recommendations through the DV4RDA initiative, beginning with a full system‑security evaluation that aligned FAIRlyz with the AIDV framework and NIST 800‑171. Findings informed targeted redesigns prioritized for policy relevance and trustworthy AI practices. Implementation by the Lifetime Omics team, supported by weekly DV4RDA feedback, strengthened authentication, improved QC integration, added per‑subject consent checks, and expanded data‑type support. A coordinated global testing effort—with structured onboarding, community outreach, cross‑time‑zone meetings, and focused subgroups—validated these updates. ImmPort datasets were prepared for FAIRlyz testing, and this combination of policy alignment, iterative development, and global collaboration enabled successful adoption of the AIDV‑WG recommendations.
Benefits and impact of adoption
Adopting AIDV‑WG policies within FAIRlyz strengthened secure biomedical data practices and demonstrated the outputs’ feasibility in real‑world use. Consent checking during QC ensured protections from the start of analysis. DV4RDA collaboration drove major improvements, including a more intuitive interface, broader data‑type and ontology support, and stricter annotation rules with validation. A full security review of FAIRlyz’s AI‑driven visitation technology also led to targeted upgrades to authentication and data protection. Key outcomes have been published:
IRB Language: https://zenodo.org/records/17703415
ICF Language: https://zenodo.org/records/17703467
Security Plan: https://zenodo.org/records/17703563
Geographies of Trust: https://zenodo.org/records/17781689
What lessons did you learn?
Strong engagement from data librarians significantly improved the data‑visitation tool, while low participation from biomedical researchers showed the need for more targeted outreach. OS variability and limited hardware reinforced the importance of running FAIRlyz on shared robust systems with group‑based access controls as it evolves toward a Trusted Research Environment. A key lesson was the value of rigorous QC: early, automated checks are essential for trustworthy analysis. The project also highlighted the need to embed data‑visitation language in IRB and consent materials and to conduct thorough security evaluations. Insights from the Geographies of Trust report underscored the need for a more intuitive workflow, motivating plans for an AI‑guided UI and a secure local AI model.
About the Adopter(s)
FAIRLYZ
FAIRlyz is a collaborative data management platform guiding organizations and their researchers to manage, share and reuse well-annotated quality data that is ready for AI. By visiting the data for AI-driven curation and quality control (QC), FAIRlyz protects sensitive information. This, along with synchronization with the central registry, forms the core functionality of the platform. FAIRlyz demonstrates broad applicability, ranging from enabling individual researchers to secure funding for data reuse to empowering organizations in strengthening internal data integrity and oversight.
Other Adoption Stories
