Understanding the implications of the US AI Executive Orders

What is an Executive Order?

EOs are official documents, numbered consecutively, through which the President of the US manages the operations of the Federal Government.  An EO is a signed, written, and published directive from the President of the US to the Federal Government. Importantly, even though EOs are binding, they must not be confused with legislation as they do not go through the legislative process (e.g., Congress). The EOs do not themselves create obligations upon business directly as they are binding only on the executive branch, e.g. federal agencies. However,  they are well-positioned to indicate the direction of future policies, as some actions can require the backing of Congress to materialize, and regulatory activities as EOs “direct” federal authorities and agencies to carry out (or refrain from carrying out)  certain courses of action. This enables EOs to provide valuable insight into understanding the areas and topics that referred authorities and agencies will act upon, for example, introducing new regulatory practices, guidelines, or courses of action.

Executive Order 13859, maintaining American leadership in artificial intelligence

The EO starts from the premise that the US is a world leader in AI research and development (R&D) and deployment and the US Government should operate with an overarching goal of continued American leadership in AI. The initiative focuses the resources of the Federal Government in support of AI innovation and does so through a coordinated strategy. 

The result is the American AI Initiative, which should be guided by five principles: (i) drive technological breakthroughs in AI across the government; ii) drive the development of technical standards; iii) the US must train its workforce with skills relevant to AI; iv) foster public trust and confidence in AI technologies through protecting liberties and privacy; v) open up markets for American AI companies.

The EO sets forth six principles that must be implemented by agencies that conduct foundational AI R&D, develop and deploy applications of AI technologies, provide educational grants, and regulate and provide guidance for applications of AI technologies. 

Notably, the EO also focuses on enhancing funding from the agencies’ budgets, granting access to computing resources, prompt standardization, direct funding to workforce programs, and creating clarity into data and model inventories held by the agencies and shall prioritize improvements to access and quality of AI data and models.

Executive Order 13960, promoting the use of trustworthy artificial intelligence in the Federal government

The backdrop of the EO is that federal agencies are encouraged to continue to use AI, but it has to be balanced with the public’s need for trust. Agencies must, therefore, design, develop, acquire, and use AI in a manner that fosters public trust and confidence while protecting privacy, civil rights, civil liberties, and American values, including adhering to all applicable laws.

Pursuant to section 3 of the EO, the following principles are put in place for the use of AI in the Federal Government:

• Lawful and respectful of American values. Agencies must use AI in a manner that respects American values and is consistent with the Constitution and other laws and regulations.
• Purposeful and performance-driven. Agencies must assess and manage the risks of using AI and ensure that the benefits of using AI outweigh the risks.
• Accurate, reliable, and effective. Agencies must ensure that the application of AI is accurate, reliable, and effective.
• Safe, secure, and resilient. Agencies must ensure that AI applications are safe, secure, and resilient.
• Understandable. Agencies must ensure that the outputs from AI applications are understandable.
• Responsible and traceable. Agencies must ensure that human responsibilities are clearly defined for AI applications and that inputs for and outputs from AI applications are documented and traceable.
• Regularly monitored. Agencies must ensure that AI applications are regularly monitored, including testing each application against these established principles.
• Transparent. Agencies must be transparent regarding their use of AI and provide relevant information to appropriate stakeholders like Congress and the public.
• Accountable. Finally, agencies shall be accountable for ensuring proper safeguards for the use and function of AI applications.

One key aspect of this executive order is the emphasis on information sharing by federal agencies regarding their use of AI. Under this EO, the Federal Chief Information Officers Council should have identified, provided guidance on, and made publicly available the criteria, format, and mechanisms for agency inventories of non-classified and non-sensitive AI use cases by agencies. Essentially, the agencies are required to prepare an inventory and thereafter share their inventories with the public while considering applicable laws, including those concerning the protection of privacy and sensitive law enforcement, national security, and other protected information.

It is important to note that although the EO’s scope regarding federal agencies is wide, it does have a significant exclusion. Namely, the Department of Defence and the Intelligence Community, and therefore intelligence AI use cases, are excluded from its application.

A clear effect of this EO is that the principles are playing a role in procurements and applications of AI, including (i) how AI applications are procured, (ii) what information will be required from contractors to provide regarding their AI applications, and (iii) how agencies test and audit government contracts for AI applications or government contracts whose performance relies on the use of AI applications. 

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

1. New standards for AI safety and security

The first action point on the EO is the most significant, as it highlights the importance of national security and requires companies to responsibly develop and utilize the most powerful or impactful AI systems. This includes life-science projects and cyber-security programs. It will be necessary to notify the government and share their safety test results and other critical information with the U.S. government before publicizing these new materials. This is a prominent step that attempts to address protecting national security and public health. Furthermore, the Biden administration wishes to develop standards, tools, and tests to help ensure that AI systems are safe, secure, and trustworthy for the military and intelligence community, private sector, and governments worldwide.

2. Protecting Americans' privacy

This section acknowledges that one of the risks of AI development is the further deterioration of the privacy of individuals around their data. How can AI systems be trained to preserve Americans' privacy better? The president calls on Congress to better protect Americans’ privacy, including from the risks posed by generative AI, and to pass bipartisan data privacy legislation to protect all Americans, with a special focus on kids. Congress also prioritizes federal agencies’ support for accelerating the development and use of privacy-preserving research. 

3. Advancing equity and civil rights

The aim is to address discrimination and civil rights violations related to AI to provide clear guidance to landlords, federal benefits programs, and federal contractors. Furthermore, calls to address algorithmic discrimination by developing training, providing technical assistance, and enhancing coordination between the Department of Justice and federal civil rights offices. Lastly, it calls to ensure fairness across the criminal justice system.

4. Standing up for consumers, patients, and students

This part of the EO speaks of two big categories harnessing the benefits of AI to promote - firstly, healthcare and, secondly, transforming education. On one hand, the priority is to enhance the American healthcare system and develop affordable and life-saving drugs. Still, no program or guidance has been given on implementing that. However, from a more practical standpoint, the Department of Health and Human Services is now tasked with developing a safety program. The aim is to address unhealthly received reports of — and act to remedy – harms or unsafe healthcare practices involving AI.

5.  Supporting workers

With this section, the President addresses one of the hottest topics for the American population - protecting workers' rights from AI harm. This clause has raised the concern that AI is changing America’s jobs and workplaces, promising improved productivity. In addition,  the dangers of increased workplace surveillance, bias, and job displacement are becoming more frequent and prominent. The intentions of the EO are clear - to put more power in the hands of the workers (organizational management), review their compensation system, and evaluate the job applications fairly. Also, the EO calls for the research on labor disruptions to be enhanced to explore their scale and nature.

6.  Promoting innovation and competition

This section is being raised because America currently positions itself as one of the leaders in AI innovation. As the EO points out, more AI startups raised first-time capital in the United States last year than in the next seven countries combined. Subsequently, the EO builds on the aim of leading the way of innovation by catalyzing AI research across the US, promoting a fair, open, and competitive AI ecosystem by “providing small developers and entrepreneurs access to technical assistance and resources, helping small businesses commercialize AI breakthroughs, and encouraging the Federal Trade Commission to exercise its authorities.”

7. Advancing American leadership abroad

As the number of regulatory challenges for AI models keeps increasing globally, this section focuses on cross-national collaboration to develop and regulate AI. The president aims to “expand bilateral, multilateral, and multi-stakeholder engagements to collaborate on AI.” To harness AI’s benefits in managing its risks and ensuring safety, the State Department, in collaboration with the Commerce Department, will work on establishing international frameworks with international partners and standards organizations.

8.  Ensuring responsible and effective government use of AI

The last section of the EO can be compared to ethical guidelines on the application of AI technology and how the US government intends to govern AI. The main action point is to enhance the government systems, which can pose risks—issuing guidance on how organizations can use AI and helping agencies acquire AI products more efficiently.

Summary

Although the Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence is still a relatively new initiative, it is already different in nature compared to the previous two EOs on AI. This is due to the fact that its detailed nature should enable it to ensure more thorough and consistent implementation. Historically, according to Stanford University’s White Paper on the Implementation Challenges to Three Pillars of America’s AI Strategy, the EOs on AI have not been successfully implemented. According to Stanford’s white paper, the reasons for this kind of high prevalence of non-implementation can be suggested to come from “a leadership vacuum and capacity gap at the agency and national level”, whereas “[A]gencies require leadership and resources and to meaningfully advance the objectives of these legal mandates''. Therefore, it is to be seen whether the latest EO will survive the test of time and be implemented successfully to carry a meaningful impact.

The information presented in this article is solely for informational purposes and should not be construed as legal or any other form of advice.