Frontiers has launched Frontiers FAIR² Data Management, presented as an all-in-one, AI-powered service designed to make research data reusable, citable, and credited.
Most scientific data do not contribute to the discoveries they could. For every 100 datasets created, around 80 remain in laboratories, 20 are shared but rarely reused, fewer than two meet FAIR standards, and only one typically drives new findings. The result includes delays in cancer treatments, climate models with evidence gaps, and research that is difficult to reproduce.
The service brings curation, compliance checks, AI-ready packaging, peer review, an interactive portal, certification, and lifetime hosting into a single workflow, with the aim of ensuring that research funded today delivers faster breakthroughs in health, sustainability, and technology. FAIR² extends the FAIR principles (Findable, Accessible, Interoperable, and Reusable) with an open specification intended to ensure that every dataset is AI-ready and responsibly reusable by humans and machines. Frontiers FAIR² Data Management is the first implementation, launched at a time when research outputs are growing and AI is reshaping discovery, turning principles into practical infrastructure at scale.
Dr Kamila Markram, co-founder and CEO of Frontiers, highlighted that a large share of scientific output risks being lost and has added that the FAIR² service is intended to ensure datasets contribute to progress while allowing contributors to receive appropriate credit.
Tasks that once required months of manual work—such as curating datasets, checking compliance, creating metadata, and preparing publishable outputs—are executed in minutes by the AI Data Steward, powered by Senscience, the Frontiers venture behind FAIR².
With one submission, researchers receive four outputs: a certified Data Package, a peer-reviewed and citable Data Article, an Interactive Data Portal with visualizations and AI chat, and a FAIR² Certificate. These components include quality checks and concise summaries to assist non-specialists and to support combining datasets across disciplines; together, they are intended to preserve, validate, cite, and enable reuse of datasets while recognizing contributors, and to make data visible and easier to explore for researchers, policymakers, practitioners, communities, and machines.
Flagship pilot datasets include SARS-CoV-2 Variant Properties, which covers 3,800 spike protein variants and links structural predictions from AlphaFold2 and ESMFold with ACE2 binding and expression data to support pandemic preparedness and analysis of variant behavior and fitness; Preclinical Brain Injury MRI, a harmonized set of 343 diffusion MRI scans from four research centers, standardized across protocols and aligned for comparability to support reproducible biomarker discovery and cross-site analysis in preclinical traumatic brain injury research; etc.
Researchers testing the pilots have noted that the FAIR² approach preserves and shares data while strengthening confidence in reuse through quality checks, clear summaries for non-specialists, and the reliability to combine datasets across disciplines, with mechanisms to ensure credit for contributors. All pilot datasets comply with the FAIR² Open Specification, with the intention that they are responsibly curated, reusable, and trusted for long-term human and machine use.
Each reuse increases the value of the original dataset, helping to avoid wasted discoveries, catalyze subsequent breakthroughs, and improve recognition for researchers. Dr Sean Hill, co-founder and CEO of Senscience, noted that substantial investments in data generation often go under-recognized and has added that FAIR² is intended to ensure citation and recognition for datasets and their creators, with the goal of accelerating advances in health, climate, and technology.
Click here to read the original press release.
More News in this Theme
