Becoming FAIR²: How a Global Research Community is Shaping the Future of AI-Ready Data
Shaped by researchers worldwide, FAIR² is evolving into a practical, community-driven standard for ethical and next-generation data sharing.
Over 400 researchers from more than 40 countries have responded to the Frontiers FAIR² Data Management Pilot, powered by Senscience’s AI platform. Whether contributing requirements, providing feedback, or hands-on testing the platform and open specification, this global community is helping define what practical, reusable, and trustworthy research data should look like in the next era of science. Selected researchers are creating and publishing data articles with AI-powered data portals and AI-ready data packages, while the broader community informs every part of the process through their requirements, feedback, and participation. In just a few months, the pilot has also attracted funders, policymakers, and organizations who see it as a catalyst for transforming the research data ecosystem.
FAIR² (FAIR Squared) is an open, standards-based spec for reusable, AI-ready research data, scheduled for release later this year. It builds on MLCommons Croissant, W3C’s JSON-LD/RDF with SHACL validation, schema.org vocabularies, and QUDT units—making datasets not just FAIR, but also machine-actionable and semantically rich. FAIR² adds Croissant RAI for ethics, governance, and bias metadata, and uses CRediT and CRO IDs for detailed contributor attribution and provenance. It enforces SHACL-based validation with actionable feedback, supports domain-specific profiles, standardizes units, enables persistent linking, and integrates with data workflows—all within a modular, community-driven roadmap that stays compatible with Croissant, schema.org, and diverse scientific domains.
To bring the specification to life, Senscience is building and operating the first implementation: the Frontiers FAIR² Data Management service. This AI-powered platform puts the FAIR² specification into practice, offering intuitive workflows for preparing datasets, metadata structuring, certification, and data article publication. At its core is Clara, an agentic AI Data Steward—transparent, researcher-guided, and built for responsible, high-quality data management.
Openness, sovereignty, and researcher control are also central: your data stays yours—never locked in or monetized. We will also support deposition into trusted repositories for maximum reusability and compliance.
Senscience is founded on the belief that research infrastructure should be shaped by its community and built on open standards—enabling a rich, FAIR²-aligned ecosystem without vendor lock-in. Launching later this year, it will establish nonprofit governance for the FAIR² specification and trademark, ensuring globally coordinated, community-led certification. As part of this effort, Senscience has joined LIFES (Leiden Initiative for FAIR and Equitable Science) to engage researchers, institutions, and data stewards worldwide—an initiative already drawing new participants into the pilot. The platform will also open to additional publishers, institutions, and organizations beyond Frontiers.
Why Researchers Are Joining
The FAIR² Data Management pilot is truly global—a diverse community brought together by shared goals and practical needs. Here’s what researchers and data stewards are saying:
Advancing open science:
“FAIR principles—ensuring data is Findable, Accessible, Interoperable, and Reusable—are essential to our mission.”
Seeking real-world solutions:
“My research data has untapped potential, but I’ve struggled with technical and time barriers.”
Building trustworthy, researcher-guided AI:
“AI-assisted curation allows researchers to focus more on discovery.”
Pursuing recognition and visibility:
“Publishing datasets with DOIs increases visibility and impact.”
Championing equity and global capacity:
“Great for building capacity and standards in underrepresented regions.”
These quotes are just a glimpse of the global voices engaged in the FAIR² pilot—scientists, data stewards, and leaders collaborating across regions to build a practical, open, and truly international research data community.
What We’re Hearing
Through interviews, walkthroughs, and direct feedback—including from funders, policymakers, and foundations—several major themes have emerged:
1. Researchers want support—not just standards
“It’s hard to know where to start, and easy to get stuck on standards that weren’t made for our kind of data.”
“Curating data is a time-consuming yet essential task… this tool could turn my research more efficient.”
“We’re using this to rethink how our department manages and shares datasets—not just one project.”
Researchers are calling for context-sensitive, workflow-driven help—rather than more checklists.
2. FAIR ≠ open
“We need solutions that allow us to describe and share our data responsibly, even when open access isn’t possible.”
“A lot of our data can’t be shared openly, but we want it to be discoverable and reusable within ethical boundaries.”
Many work with sensitive, restricted, or ethically governed data. They’re asking: how can we make data findable and reusable—even when it can’t be fully public?
3. Metadata must accommodate complexity
“Our dataset includes structured clinical data, free-text notes, and coded survey responses—we struggle to describe it all in one metadata format.”
Mixed-methods, multilingual, and interdisciplinary datasets require flexible, layered metadata—not rigid templates.
4. AI should support—not replace—researchers
“I’m excited by the potential of AI to support—not replace—human expertise in research data management.”
“If I can spend less time wrangling metadata, I can spend more time analyzing and collaborating.”
Researchers see AI as a promising partner—but insist it must remain transparent, editable, and under their control.
5. Ecosystem Momentum
Funders, policymakers, and foundations are joining the conversation—recognizing FAIR² as a catalyst for change across the research data landscape, and a way to incentivize, support, and monitor responsible data sharing at scale.
“We see FAIR² as an opportunity to raise the bar for responsible, transparent research data management.”
“Having a validated, AI-ready standard with an accompanying publication could help us incentivize best practices and track compliance more effectively.”
“This is a chance to align funding, policy, and technology around data that can be trusted and reused.”
Who’s Involved & What We’re Exploring
Researchers from Europe, North and South America, Africa, Australia, and Asia are contributing to the FAIR² pilot. The community includes early-career and senior scientists, data stewards, librarians, and infrastructure teams, all eager to explore FAIR²’s potential across real research contexts.
Researchers are interested in applying FAIR² to a remarkable variety of datasets, including:
Biomedical and Clinical: Patient registries, clinical trial data, molecular markers, immunology, rare disease cohorts, health systems, maternal/child health, and epidemiological studies
Genomics and Multi-omics: Genotyping, single-cell and bulk RNAseq, metagenomic, and metabolomics data, often paired with clinical or environmental measures
Environmental and Earth Sciences: Climate and weather model outputs, oceanographic and freshwater observations, air pollution, and environmental monitoring datasets
Social and Behavioral Science: Longitudinal surveys, educational research, psychological assessments, and public health datasets from around the world
Computational Science and Engineering: Simulation outputs (climate, tsunami, Monte Carlo), digital elevation models, sensor/process engineering data, and large-scale image data
Agriculture, Biodiversity, and Life Sciences: Plant phenotyping, agricultural monitoring, biodiversity collections, and microbial communities
Policy, Economic, and Administrative Data: Open government data, legal documents, economic indices, and policy datasets
Other: Neuroimaging, neuroscience, proteomics, biocatalysis, pain research, and complex, multi-modal datasets that combine quantitative and qualitative or structured and unstructured sources
Data formats span spreadsheets, CSV, JSON, image and video files, geospatial formats, and scientific standards. Many are particularly interested in how FAIR² can help with privacy-preserving sharing, metadata standardization, FAIRification of legacy data, and supporting open science, AI-readiness, and citation.
Through the pilot, we’re testing and refining FAIR² to address a wide range of data types—from biomedical and genomics to environmental, computational, social, and multi-modal datasets—and how evolving community needs may guide future support and development.
We look forward to sharing updates as new datasets are published.
What’s Coming Next
As we approach launch, our focus is on outputs, improvement, and deeper engagement:
Preparing the next set of FAIR² Data Articles, Packages and Portals for peer review and publication
Applying what we’re learning to improve metadata flows, usability, and researcher experience
Continuing to train and expand the FAIR² AI Data Steward, making it easier and more intuitive for researchers to achieve best practices without getting stuck on manual curation
Expanding outreach through expanded user testing, co-design workshops, discipline-specific guidance, and community support
Testing, refining, and preparing the Frontiers FAIR² Data Management service for launch later this year
This next phase turns insight into infrastructure—delivering tools that reflect the real needs and ambitions of a global, cross-sector research community.
Want to Join the FAIR² Community?
Whether you’re working with legacy data, creating a new dataset for publication, supporting colleagues, or helping define data policy, you can play a role in the FAIR² community.
🔗 Learn more about Frontiers FAIR² Data Management or request to get involved
FAIR² is about building practical, trusted solutions for how science actually happens—with the whole research community, not just for it.
We welcome feedback, collaboration, and new perspectives as we continue building—please get in touch.