The FAIR Data Innovations Hub Is a Part of a New National Artificial Intelligence (AI) Initiative!

The FAIR Data Innovations Institute will play an important role in the National Institutes of Health (NIH)’s new initiative to unleash the full potential of artificial intelligence (AI) for improving medical care.

The NIH Bridge2AI program

The NIH has announced the launch of the Bridge2AI program. The program’s goal is to help set the stage for the wider use of AI to solve pressing challenges in human health. Biomedical research and health care already use AI, but its widespread adoption has been limited by the quality of the data available. Indeed, to use AI to its fullest potential, the data given to AI models need to include important contextual information about the data type, collection conditions, and other parameters. In addition, careful attention must be paid to the social and ethical contexts in which the data is collected to help prevent the introduction of bias or inequities in the AI models. Unfortunately, the data available today typically lack contextual information and are not adequately representative of the diverse populations.

Setting the benchmark for AI-ready datasets

Bridge2AI has issued awards to four Data Generation projects whose work over the next four years will set the foundation of the program by generating new data sets that are ready for AI systems. In addition, these projects will establish new tools and standards for making data AI-ready as well as develop training materials that promote a culture of diversity and the use of ethical practices throughout the data generation process. The outcomes of the Data Generation projects will be made widely available so other biomedical research projects can easily replicate their approaches to generate and disseminate AI-ready data sets.

The AI-READI project

The FAIR Data Innovations Hub will co-lead tools development for one of the Data Generation projects called Artificial Intelligence Ready and Equitable Atas for Diabetes Insights (AI-READI). Using diabetes as its model disease, the project will aim to collect data from 4,000 participants. To ensure the data is population-representative, the 4,000 participants will be balanced for three factors: disease severity, race/ethnicity, and sex. Various data types will be collected from each participant, including vitals, electrocardiogram, glucose monitoring, physical activity, ophthalmic evaluation, etc. The FAIR Data Innovations Hub will co-lead the development of a cloud platform called fairhub.io that will allow data-generating researchers to collect, prepare, and share their data with ease. Our team will focus on integrating tools into fairhub.io that will assist these researchers to make their data compliant with the Findable, Accessible, Interoperable, Reusable (FAIR) data principles, which are widely adopted guidelines for optimizing the reuse of data by humans and AI systems. The project will benefit from a collaboration with Microsoft for developing, hosting, and maintaining fairhub.io through Microsoft’s cloud computing service Azure.

Funding and collaborators

This project is supported by the National Institutes of Health (OT2OD032644). The AI-READI team will receive $7.8 million in funding for the first year, distributed across collaborating institutions, to kick off the project. In addition to the FAIR Data Innovations Hub, other institutions collaborating on the AI-READI Data Generation project include: University of Washington, Oregon Health & Science University, Johns Hopkins University, University of California at San Diego, University of Pennsylvania, Stanford University, Native BioData Consortium, University of Alabama at Birmingham, and Microsoft.

We are excited!

The FAIR Data Innovations Hub is delighted to be part of this new NIH initiative where we will continue our mission of building open source tools that make it easier for researchers to prepare and share FAIR data.