AI Commons Consortium (AICC)

The AI Commons Consortium (AICC) is a consortium managed by the 501(c)(3) Center for Computational Science (CCSR) with the goal of helping to develop and support open-source AI Commons and AI Meshes based upon the open-source Gen3 data platform. Gen3 follows GA4GH and other relevant standards.

One of the main projects for the AICC is to develop the Genomic AI Commons (GAC), which is a Gen3 AI Commons for cancer genomics that leverages data from the Genomic Data Commons (GDC). The GDC is the world’s largest collection of harmonized cancer genomics datasets in the world and is used by over 175,000 researchers each month. 

With your own Gen3 AI Commons (or another data commons with compatible services), you can leverage information from the GAC and participate in federated learning and federated inference, without any of your data leaving your commons.

The Gen3 AI Commons are based upon the technology developed by the Meshes of Midscale Models Initiative (M3) led by the Center for Translational Data Science at the University of Chicago.  The M3 architecture is based upon a core set of AI services that are supported by Gen3 AI Commons, which includes a model repository, agentic services, and services for embeddings, inferencing, and training small and midscale AI models.  With these AI services, data does not have to leave a Gen3 Commons and you can use AI models to search, discover and explore your data. Of course, you can also interoperate with third party models, but M3 studies have shown that with high quality data, you can build your own small and midscale models which in many cases are as powerful or more powerful than Frontier models.  

The M3 Initiative is also integrating services into Gen3 Commons and Gen3 Meshes so two or more Gen3 AI Commons can support federated learning and federated inferencing and build more powerful AI models without any data leaving a Gen3 AI Commons in the AI Mesh.  The technology is structured so that a Gen3 AI Commons can easily join multiple AI Meshes and interoperate in a trusted manner using what are called node and mesh cards (an emerging GA4GH standard).

The AICC is structured like the BLOODPAC consortium, with memberships based upon the size of your organization and all work done through AICC Working Groups.  Benefits of AICC membership include:

  • Early access to Gen3 AI Commons and Gen3 AI Meshes technology.

  • The ability to shape the priorities of the AICC.

  • Providing support to accelerate the development of the Genomic AI Commons and help improve the outcomes of cancer patients. 

  • Gain practical experience with small and midscale AI models and associated agentic services and learn how they can complement and interoperate with third party models, including frontier models.

Please reach out to us at info@occ-data.org to learn more about the AICC.