OMB: How Can We Enhance Federal Data to Boost AI Research?

The White House wants public insight on the federal datasets and models it can release or improve to invigorate artificial intelligence research and testing efforts, according to a request for information released by the Office of Management and Budget Thursday. 

“The Trump administration understands that unleashing federal data and models is critical to drive top-notch AI research and promote technological breakthroughs and competitiveness,” Michael Kratsios, the president’s deputy assistant for technology policy, said in a statement about the request. “[W]e also take seriously the need to balance access to data and maintaining the civil liberty and privacy protections Americans expect.”

As the latest piece of the administration’s national strategy to boost commitment to global leadership in AI, the RFI is an extension of an executive order Trump signed in February: the American AI Initiative. 

The request notes that many federal datasets are accessible through data.gov, whether they are fully available to the public or obtained with restricted use. But OMB says this information “may or may not” be suitable for AI research, motivating the need for public feedback. 

OMB asks individuals to elaborate on the federal data and models they access, as well as restricted and private datasets agencies could open up to promote AI R&D. The agency engages the public to provide details around technical issues users may face when working with the data and key gaps in federal model availability that’s hampering AI innovation. It also asks about the questions and problems users are seeking to answer through implementing the data and deploying the tech and what data sharing considerations the government should include in federally-funded agreements. 

The request also includes questions to help agencies better understand how their data can be utilized. 

“As agencies review their data and models, what are the most important characteristics they should consider?” it asks. “Stated differently, what characteristics of data sets or models make them well-suited for AI R&D?”

The agency also notes that characterizations of certain terms can be flexible. 

“Respondents to this RFI may define ‘data,’ ‘data set,’ and ‘model’ as they desire, indicating clearly what they mean when using the term,” it said.

The agency set the deadline for 30 days after the RFI was released, so submissions will be open until Aug. 9. OMB said anonymous responses will be accepted, but it also asks respondents to share their missions or areas of expertise. 

Earlier this month, the National Institute of Standards and Technology published a draft version of its plan for federal engagement around developing technical standards to promote U.S. leadership in AI, marking another effort directed by the national AI initiative.