Commerce seeks insight on how to make its data more AI-friendly

monsitj/Getty Images

A new request for information issued by the Department of Commerce asks for expert commentary on how the agency can better structure and disseminate its publicly available data.

The U.S. Department of Commerce wants to make its open dataset reserves better primed to help train generative artificial intelligence systems and further democratize access to technological resources needed to create trustworthy and reliable AI and machine learning systems. 

Outlined in a request for information published on April 17, Commerce is asking for comments in the form of recommendations and suggestions from industry experts on how to best structure and distribute the public data produced by the agency. 

The ultimate goal of the RFI is to provide easy public access to high-quality data.  

“AI tools are increasingly used for data analysis and data access, so Commerce hopes to ensure that the data these tools consume is easily accessible and ‘machine understandable,’ versus just ‘machine readable,’ the RFI reads. “Therefore, this RFI explores how to achieve better data integrity, accessibility, and quality for emerging AI technologies.”

More technical details Commerce wants to explore focus on connecting human terms to data elements via metadata knowledge graphs and harmonizing standardized ontologies to Commerce’s internal vocabularies. 

The RFI particularly calls for insight on embracing ontology, with potential AI needs. In the context of AI, new research approaches are working on incorporating ontological knowledge graphs into AI system operations, including benchmarking output results and seamlessly incorporating external data. 

Developing and using metadata standards has also been a priority item for Commerce’s AI endeavors. Metadata is often used within relational databases and other systems to execute commands in software programs. The RFI also asks for feedback on using metadata when programming AI models.

Commerce is working to execute several items mandated under President Joe Biden’s 2023 Executive Order on AI, with democratizing crucial components to building safe AI systems, such as training datasets, being a priority.

The RFI comment period closes on July 16. 

Earlier this year, the National Telecommunications and Information Administration, a sub agency of Commerce, sought public comment on open-weight AI systems to help craft policy and innovation.