DoD Responsible AI (ethical principles and AIA Toolkit)

Naming note

Several official sites now use “Department of War” for the Department of Defense. Older documents say DoD. This site uses the name that appears in each source.

The five AI ethical principles

The Department of Defense officially adopted a series of ethical principles for the use of AI in February 2020, following recommendations from the Defense Innovation Board. They apply to how Department personnel develop, deploy, and use AI capabilities. The five principles are:

1. Responsible - personnel will exercise appropriate levels of judgment and care while remaining responsible for the development, deployment, and use of AI capabilities.
2. Equitable - deliberate steps to minimize unintended bias.
3. Traceable - capabilities developed and deployed so that personnel understand the technology, including transparent and auditable methodologies, data sources, and design procedures and documentation.
4. Reliable - explicit, well-defined uses, with safety, security and effectiveness subject to testing and assurance within those uses across their entire life-cycles.
5. Governable - the ability to detect and avoid unintended consequences, and to disengage or deactivate deployed systems that demonstrate unintended behavior.

Strategy and Implementation Pathway

The RAI Strategy and Implementation Pathway was signed in June 2022 by Deputy Secretary of Defense Kathleen Hicks and spells out 64 lines of effort for operationalizing the principles.

The toolkit (RAI, now titled AIA)

The Chief Digital and Artificial Intelligence Office (CDAO) publicly released the RAI Toolkit on 14 November 2023. That announcement said it provides a voluntary process that identifies, tracks, and improves alignment of AI projects to RAI best practices and the DoD AI Ethical Principles. It is built upon the DIU Responsible AI Guidelines and Worksheets, the NIST AI Risk Management Framework and Toolkit, and the IEEE 7000 standard, among other sources.

The live Tradewinds page now titles the same public resource the AI Assurance (AIA) Toolkit. The older RAI Toolkit URL still resolves. It remains a self-evaluation resource that helps teams assess their AI projects at any stage of development, ensuring alignment with DoW AI Ethical Principles, and is accessible in both IL2 and IL5 environments. The 2023 Tradewinds copy described the release as a minimum viable product; the AIA page we fetched on 14 October 2026 does not use that label.

What it is not

The toolkit is a voluntary assessment resource. We have not found a source that makes completing it a contractual or certification requirement, and we make no such claim. For handling controlled data, see CMMC and NIST SP 800-171; for the NIST framework it builds on, see NIST AI RMF. UK MOD direction is JSP 936; the Allied baseline is NATO’s Principles of Responsible Use. The directory lists the CDAO AIA Toolkit as a government-published resource. Tool selection for red-teaming is Choosing AI red-teaming tools for safety-critical programs.

The Department’s release also says the toolkit offers guidance and an RAI standard to current and future DoD industry partners. Check the toolkit itself for what that means for a specific contract.