This Executive Order, issued on December 3, 2020, by President Donald J. Trump, outlines policies and principles for the responsible use of artificial intelligence (AI) within the U.S. Federal Government. Key points include:
- Purpose:
- AI is recognized as a tool to enhance government efficiency, effectiveness, and decision-making while improving public trust and adherence to U.S. values.
- Encourages agencies to use AI responsibly to address tasks such as fraud prevention, regulatory compliance, predictive maintenance, and more.
- Policy Goals:
- Promote AI innovation to improve government operations.
- Ensure AI use aligns with privacy, civil rights, and national values.
- Build public trust in AI by adhering to principles of transparency, accountability, and safety.
- Principles for AI Use:
- Lawful and Respectful: AI must align with the Constitution, laws, and U.S. values.
- Purposeful: Benefits must outweigh risks, which should be assessed and managed.
- Reliable and Effective: Ensure accuracy and consistency with intended use cases.
- Safe and Secure: Protect AI against vulnerabilities and adversarial threats.
- Understandable: Maintain clarity for users and stakeholders.
- Responsible: Clearly define roles and ensure traceability of AI applications.
- Transparent: Provide disclosures about AI use where applicable.
- Accountable: Agencies must monitor, audit, and ensure proper use of AI.
- Implementation:
- The Office of Management and Budget (OMB) will issue guidance for implementing these principles.
- Agencies must inventory AI use cases, review for compliance, and either update or retire non-compliant applications.
- Interagency Coordination and Expertise:
- Agencies are directed to collaborate through interagency bodies and recruit AI expertise via programs like the Presidential Innovation Fellows.
- Training programs and rotational opportunities will expand AI knowledge within the Federal workforce.
- Scope:
- Applies to AI used in non-defense and non-intelligence contexts for mission-critical applications.
- Excludes common commercial AI products, defense/national security AI systems, and AI R&D activities.
