Responsible

Humans (DOD personnel) are responsible for the development, deployment, and use of AI

Equitable

Unintended bias in AI capabilities must be minimized

Traceable

Humans (relevant personnel) will possess an appropriate understanding of the technology, development processes, and operational methods applicable to AI. AI must have auditable methodologies, data sources, and design procedures and documentation

Reliable

AI must have well-defined uses. The safety, security, and effectiveness of AI will be subject to testing across their entire life cycle

Governable

AI will possess the ability to detect and avoid unintended consequences. Humans will have the tools to deactivate AI that demonstrates unintended behavior