Pentagon’s CDAO Issues Toolkit on Responsible AI Best Practices

The Department of Defense Chief Digital and Artificial Intelligence Office has issued the Responsible Artificial Intelligence Toolkit, one of the drivers outlined in the DOD’s RAI Strategy and Implementation Pathway that Deputy Secretary of Defense Kathleen Hicks signed in June 2022. The RAI S&I Pathway operationalizes the DOD’s AI Ethical Principles defining 64 tracks for responsible AI use.  

According to a DOD press release, the toolkit is a voluntary process for users to identify, track and improve AI projects’ alignment to RAI best practices. 

The toolkit features intuitive flow guides enabling users to access tools, such as customizable and modular assessments, throughout an AI project, the DOD said Tuesday. 

An RAI standard and guidance are also provided in the toolkit for the DOD’s current and future partners, it added.

The toolkit was developed based on the National Institute of Standards and Technology’s AI Risk Management Framework and the Defense Innovation Unit’s Responsible AI Guidelines and Worksheets, among others.

The DOD’s RAI Division led the toolkit’s development, with the support of the department’s Responsible AI Working Council, DOD resource persons, federal-funded R&D centers, the academe and industry partners.