Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence

Section 1: Purpose

Today, President Biden issued a landmark Executive Order to position the United States as a global leader in seizing AI’s potential while managing its risks. This Executive Order establishes new standards for AI safety, protects Americans' privacy, advances equity and civil rights, safeguards consumers and workers, promotes innovation and competition, and strengthens U.S. leadership abroad. It builds on the Biden-Harris Administration’s ongoing efforts to ensure responsible innovation, including securing voluntary commitments from 15 leading AI companies to develop safe, secure, and trustworthy AI systems.

The Executive Order directs the most comprehensive set of actions ever taken by any government to manage AI’s opportunities and challenges, ensuring AI benefits society while preventing harm.

Section 2: Policy and Principles

The Administration is committed to the responsible development of AI guided by these principles:

• (a) Safety and Security: AI systems will undergo rigorous evaluation, including red-teaming and safety testing, to mitigate risks in biotechnology, cybersecurity, and critical infrastructure. New standards will protect against AI-enabled biological threats and malicious cyber activities.
• (b) Responsible Innovation and Competition: Promoting AI research and education while ensuring fair competition prevents dominant firms from stifling small developers and entrepreneurs. The National AI Research Resource pilot will expand AI research across the U.S., and initiatives will support small businesses commercializing AI breakthroughs.
• (c) Support for Workers: AI must empower, not undermine, workers. The Executive Order mandates the development of best practices to protect workers’ rights, promote job security, and address AI’s impact on labor markets.
• (d) Equity and Civil Rights: To prevent AI from exacerbating bias and discrimination, the Administration will enforce civil rights protections, provide guidance to federal contractors, and promote fairness in AI applications across housing, justice, and healthcare.
• (e) Consumer Protection: Federal agencies will establish safeguards to protect consumers from AI-related harms, particularly in healthcare, finance, and education, while advancing AI’s potential to enhance these sectors.
• (f) Privacy and Civil Liberties: New guidelines will strengthen privacy protections by promoting privacy-preserving technologies (PETs) and addressing the risks of commercially available data used in AI systems.
• (g) Federal Government Leadership: Federal agencies will develop AI expertise, modernize technology infrastructure, and ensure AI adoption aligns with safety and ethical standards. Rapid hiring AI professionals and government-wide training initiatives will enhance federal AI capacity.
• (h) Global Leadership: The U.S. will lead international efforts to establish AI governance frameworks, working with allies and international organizations to promote safe and equitable AI deployment worldwide.

Section 3: Definitions

Key terms include:

• AI System: Any system, tool, or application that uses AI to operate.
• AI Model: A machine-learning component generating outputs.
• Generative AI: AI models producing synthetic content (images, videos, text, etc.).
• AI Red-Teaming: Structured testing to identify flaws and vulnerabilities in AI systems.
• Privacy-Enhancing Technology (PET): Tools mitigating privacy risks through secure data processing techniques.
• Dual-Use Foundation Model: Broad-use AI models that could pose security risks if misused.
• Synthetic Content: AI-generated or modified information.
• Watermarking: Embedding information into AI-generated content to verify authenticity.

Section 4: Ensuring AI Safety and Security

4.1 New Standards for AI Safety and Security

President Biden’s Executive Order establishes the most comprehensive AI safety standards ever implemented:

• Developer Reporting: Developers of robust AI systems must share safety test results and critical information with the U.S. government. Companies developing foundation models that pose risks to national security or public safety must notify the government and report the results of AI red-team safety tests.
• Testing and Benchmarking: The National Institute of Standards and Technology (NIST) will develop rigorous standards for AI safety testing. DHS will apply these standards to critical infrastructure, establishing the AI Safety and Security Board.

4.2 Preventing Biological and Cyber Threats

The Departments of Energy and Homeland Security will address AI’s threats to nuclear, biological, chemical, and cybersecurity systems by establishing strong standards for biological synthesis screening and securing critical infrastructure.

4.3 Mitigating Synthetic Content Risks

To protect against AI-generated misinformation and fraud, the Department of Commerce will develop content authentication and watermarking guidelines. Federal agencies will adopt these practices to ensure communications remain trustworthy.

4.4 Developing a National Security Memorandum on AI

A National Security Memorandum, directed by the National Security Council and White House Chief of Staff, will outline AI governance for national security, ensuring safe and ethical AI use in defense, intelligence, and cybersecurity operations while countering adversaries’ AI-enabled threats.

4.5 Protecting Critical Infrastructure

In coordination with DHS and CISA, agencies overseeing critical infrastructure will assess AI-related risks and update safety protocols. A new AI Safety and Security Board will advise on improving incident response and security standards.

Section 5: Promoting Innovation and Competition

5.1 Catalyzing AI Research and Innovation

The Executive Order launches a pilot for the National AI Research Resource, which will provide researchers and students with access to AI resources and data. Federal grants will support AI research in healthcare, climate change, and other vital areas.

5.2 Promoting Fair AI Competition

To prevent monopolistic practices and promote fair AI competition, the Federal Trade Commission (FTC) is encouraged to exercise its authority under the Federal Trade Commission Act to ensure a fair marketplace.

5.3 Supporting Small Businesses and Entrepreneurs

The Small Business Administration (SBA) will:

• Establish Small Business AI Innovation and Commercialization Institutes.
• Provide funding and technical assistance to startups.
• Revise eligibility criteria to improve small business access to AI-related resources.

Section 6: Supporting Workers

6.1 Mitigating AI’s Impact on Labor Markets

Within 180 days, the Council of Economic Advisers and the Secretary of Labor will report on AI’s labor-market impacts and recommend strategies to protect workers.

6.2 Developing Best Practices for AI in Workplaces

The Secretary of Labor will develop principles and best practices to prevent AI-related harms in workplaces, including addressing job displacement, ensuring fair compensation, and safeguarding workers' rights.

6.3 Building a Diverse AI-Ready Workforce

The National Science Foundation (NSF) will prioritize funding for AI-related education, training, and workforce development to create equitable pathways into AI careers.

Section 7: Advancing Equity and Civil Rights

7.1 Addressing Algorithmic Discrimination

To combat AI-driven discrimination, the Department of Justice and federal civil rights offices will provide training and technical assistance and ensure best practices for investigating and prosecuting civil rights violations related to AI.

7.2 Ensuring Fairness in AI Use by Government Programs

Agencies administering federal benefits will follow clear guidance to prevent AI algorithms from exacerbating discrimination and bias in their decision-making processes.

Section 8: Protecting Consumers, Patients, and Students

8.1 Strengthening Consumer Protections

Federal agencies will safeguard consumers by developing regulations that address AI’s impact on critical sectors such as healthcare, finance, and education.

8.2 Promoting Responsible AI Use in Healthcare and Education

The Department of Health and Human Services will establish a safety program to mitigate AI-related risks in healthcare and ensure responsible AI use in developing life-saving drugs. AI-powered educational tools will be guided by best practices designed for schools and educators.

Section 9: Protecting Privacy

9.1 Strengthening Privacy Protections with AI

To safeguard privacy, federal agencies will:

• Promote privacy-preserving technologies (PETs).
• Evaluate commercially available information procured by agencies.
• Develop privacy guidelines to ensure AI systems protect personal data.

9.2 Advancing Privacy Research and Innovation

The NSF will fund a Research Coordination Network to advance privacy technologies, encouraging the adoption of cutting-edge privacy-preserving solutions.

Section 10: Advancing Federal Government Use of AI

10.1 Issuing AI Governance and Oversight Guidelines

The Director of OMB will coordinate federal AI use through an interagency council, issuing guidelines to strengthen AI adoption, promote transparency, and ensure compliance with AI safety standards.

10.2 Enhancing Federal AI Capacity

Federal agencies will:

• Designate Chief AI Officers to oversee AI governance.
• Develop AI strategies and Governance Boards to monitor AI risks.
• Launch a government-wide AI talent surge to recruit and train AI professionals.

Section 11: Strengthening U.S. Leadership in AI

11.1 Promoting Global AI Standards and Partnerships

The State and Commerce Departments will lead efforts to develop international AI standards and frameworks through collaboration with allies and multistakeholder organizations.

11.2 Expanding Global AI Collaboration

The U.S. will build on existing partnerships with over 20 nations, supporting Japan’s leadership of the G7 Hiroshima Process, the UK Summit on AI Safety, and India’s Global Partnership on AI.

Section 12: Implementation

A White House AI Council, chaired by the Deputy Chief of Staff for Policy, will oversee the execution of this Executive Order. This council will ensure that AI advances U.S. security, equity, and global leadership while managing potential risks.