DIU Issues Step-by-Step Guidelines for Defense Stakeholders to Responsibly Use AI
Client Alert | 2 min read | 12.10.21
On November 15, 2021, the Defense Innovation Unit (DIU) released its Responsible Artificial Intelligence (AI) Guidelines (“RAI Guidelines”) to help contractors and federal officials gauge whether AI technology and programs align with the Department of Defense’s (DoD) Ethical Principles for AI (as we previously reported on here). Specifically, the RAI Guidelines provide a process for all stakeholders involved in AI development (e.g., program managers, commercial vendors, and government partners) to achieve the following goals:
(1) Ensure that the DoD’s Ethical Principles for AI are integrated into all three phases of the technical lifecycle (i.e., planning, development, and deployment);
(2) Effectively examine, test, and validate that all programs and prototypes align with DoD’s Ethical Principles for AI; and
(3) Leverage a process that is reliable, replicable, and scalable across a variety of programs.
The step-by-step process, which is presented as worksheets in the RAI Guidelines, instructs stakeholders on how to properly scope AI problem statements. They also provide detailed guidance on the considerations that stakeholders should keep in mind as they proceed through the planning, development, and deployment phases of AI systems. Those considerations include procedures for identifying who might use the technology, as well as who might be harmed by it, what those harms might be, and how they might be avoided – both before the system is built and after it is deployed.
As the RAI Guidelines note, DIU is the only DoD organization exclusively focused on accelerating the adoption of existing commercial technology across the U.S. military and partners with DoD organizations and “non-traditional” companies to field commercial solutions. According to DIU, although the majority of AI companies it works with are new to doing business with the Federal government, generally they are open to doing business with DIU because of its more agile and “commercial friendly” acquisition processes. Specifically, DIU utilizes other transaction (OT) authority under 10 U.S.C. § 2371b to direct-award prototype agreements, and successful prototypes may result in the direct award of a sole-source follow-on production OT or contract.
According to DIU, it is “actively deploying the RAI Guidelines on a range of projects that cover applications including predictive health, underwater autonomy, predictive maintenance, and supply chain analysis.” Additionally, DIU anticipates continued collaboration with government, industry, academia, and civil society to further develop the RAI Guidelines, and invites stakeholders to provide feedback, comments, or suggested updates.
Crowell & Moring is available to assist companies of all sizes and in all markets navigate the transition from commercial contracting to government contracting efficiently and effectively.
Contacts

Partner and Crowell Global Advisors Senior Director
- Washington, D.C.
- D | +1.202.624.2698
- Washington, D.C. (CGA)
- D | +1 202.624.2500
Insights
Client Alert | 3 min read | 06.12.26
DOJ Guidance Backs Away From Disparate Impact Liability
On June 9, 2026, the U.S. Department of Justice (DOJ) issued a formal opinion concluding that the Equal Opportunity Employment Commission’s (EEOC) existing interpretations of Title VII of the Civil Rights Act of 1964 (Title VII) disparate-impact liability, including the Uniform Guidelines on Employee Selection Procedures (UGESP), are unconstitutional. According to the opinion, EEOC’s prior interpretations contemplate liability based on disproportionately adverse effects alone, without regard to an employer’s likely intent, rather than treating disparate impact as an evidentiary mechanism to “smoke out” intentional discrimination. DOJ found that this approach functions as a “qualified racial-proportionality mandate” that places “a racial thumb on the scales, often requiring employers to evaluate the racial outcomes of their policies, and to make decisions based on (because of) those racial outcomes.” The opinion fulfills one mandate of Executive Order 14281, which rejected disparate-impact liability insofar as it “creates a near insurmountable presumption that unlawful discrimination exists wherever there are any differences in outcomes among different [demographic groups].”
Client Alert | 4 min read | 06.12.26
Auto Dealers: The FTC Is Back in the Driver’s Seat — Warning Letters Signal Renewed Federal Scrutiny
Client Alert | 13 min read | 06.12.26
Client Alert | 4 min read | 06.12.26




