Introduction

The one-year anniversary of President Biden’s Executive Order on artificial intelligence (AI) marks a pivotal moment to assess its impact, especially considering potential administrative changes. This Executive Order [1] [3] [4] [5] [6] [7] [8] [9] [10], the most extensive in history, outlines numerous actions for federal agencies to manage AI’s risks and opportunities across various sectors, including healthcare [1] [5] [8] [9], education [2] [3] [4] [8] [9], and international collaboration [9].

Description

The one-year anniversary of President Biden’s Executive Order (EO) on artificial intelligence (AI), issued on October 30, 2023 [9] [10], presents a critical opportunity to evaluate its impact [8], particularly in light of potential changes in administration [8]. By October 24, 2024 [1] [9], federal agencies had successfully completed over one hundred mandated actions outlined in the EO, which is the longest in history at 110 pages and encompasses a wide range of areas affected by AI, including healthcare [1] [5] [8] [9], international collaboration [9], education [2] [3] [4] [8] [9], and workforce implications [9].

The Biden-Harris Administration has made significant strides in managing the risks and harnessing the potential of AI over the past year [2]. The EO establishes new standards for AI safety and security [5], promotes equity and civil rights [5] [6], safeguards consumers and workers [5] [10], and encourages innovation and competition [5] [10]. It outlines extensive actions for federal agencies regarding AI policy and usage [6], with 150 requirements across various sectors, including employment [6], housing [2] [3] [4] [6] [8] [9], and civil rights [2] [3] [4] [5] [6] [10]. Key actions include the establishment of the National Institute of Standards and Technology’s (NIST) AI Safety Institute (US AISI) [10], which serves as the primary point of contact for industry within the government [10], providing critical guidance and conducting pre-deployment testing of frontier AI models to identify vulnerabilities and enhance resilience. This initiative is part of a broader effort to ensure that AI practices do not infringe on rights and that public funds are utilized effectively [10]. The EO also led to the appointment of chief AI officers (CAIO) within federal agencies to enhance oversight and governance.

The first National Security Memorandum (NSM) on AI was issued [7] [8], promoting federal leadership in global AI development and establishing governance frameworks [9]. This memorandum directs federal agencies to ensure the safe and trustworthy development of AI while promoting human rights and democratic values [2]. It emphasizes the importance of Privacy Enhancing Technologies (PETs) to protect sensitive data used in AI training and establishes enhanced safety protocols and reporting requirements for developers of powerful AI systems, including a proposed rule requiring companies to report critical safety test results and training plans [7].

The Office of Management and Budget (OMB) has mandated federal agencies to document their risk mitigation strategies regarding AI systems that could affect public safety and civil rights [3] [4]. OMB has also provided guidance on responsible procurement practices for AI [3] [4], emphasizing the need for third-party vendors to implement necessary safeguards [3] [4]. The Department of State has published the Global AI Research Agenda [8] [9], emphasizing principles such as inclusion [8] [9], responsible research conduct [8] [9], and collaboration [7] [8] [9], while prioritizing sociotechnical research and AI safety [8] [9]. Additionally, the AI in Global Development Playbook outlines the challenges and opportunities of leveraging AI in low- and middle-income countries [9], focusing on sustainable development and human rights [9].

In healthcare [2] [8] [9], the Department of Health and Human Services has initiated an AI Safety Program to monitor and mitigate risks associated with AI [8] [9], ensuring compliance with civil rights regulations [2]. The Department of Labor has introduced the AI & Inclusive Hiring Framework to enhance accessibility in AI-enabled hiring processes [8] [9], ensuring compliance with legal standards and protecting workers in AI-related environments. The Department of Labor has also introduced best practices for AI governance in the workplace [3], focusing on protecting workers’ rights and ensuring proper use of worker data [3]. The Department of Housing and Urban Development has clarified that housing providers must guard against discrimination in AI-driven rental decisions [3] [4], addressing potential biases in housing practices [4].

The Department of Education has released a toolkit aimed at ensuring equitable AI integration in K-12 schools [3] [4], reinforcing existing civil rights laws and guiding leaders in the ethical integration of AI in classrooms. The Office of Management and Budget has provided guidance on the responsible acquisition of AI in government [8], emphasizing cross-agency collaboration and risk mitigation during procurement [8].

Furthermore, the Committee on House Administration has urged the US Copyright Office to release pending reports on AI’s implications for copyright law [8] [9], following the initial reports published earlier in 2024 [8]. Legal frameworks have been reinforced to ensure non-discriminatory practices in housing and public benefits programs [2], with the Department of Justice coordinating civil rights efforts related to AI [2]. However, concerns remain regarding the enforcement mechanisms for civil rights protections within complex AI systems, necessitating clear accountability measures from Congress.

In 2023, the Treasury’s Financial Stability and Oversight Committee (FSOC) convened a symposium on AI and financial stability [6], highlighting the need for a comprehensive understanding of AI’s risks and opportunities in the financial sector [6]. Secretary Yellen announced a request for information (RFI) aimed at gathering insights from various stakeholders [6], including financial institutions and technology experts [6], to inform regulatory approaches [6].

The Department of Commerce has established the US AI Safety Institute and formed a consortium of approximately 280 members focused on AI safety [7]. New guidance and software have been released to enhance the safety and security of AI systems [7], and formal collaborations with leading AI companies have been initiated for research and testing [7]. Voluntary frameworks for managing risks associated with generative AI and dual-use foundation models have also been published [7], alongside a Request for Information on responsible AI development in chemical and biological sciences [7].

A final report on standards for addressing risks from AI-generated content has been submitted to the White House [7], and a challenge has been launched to develop detection methods for such content [7]. The Department has also released a report on dual-use foundation models [7], including policy recommendations based on extensive stakeholder outreach [7]. Additionally, a competition has been announced to support AI-enabled research in sustainable semiconductor materials [7], and the US Patent and Trademark Office has issued guidance on the patentability of AI-assisted inventions [7].

International engagement on AI safety has been prioritized [7], with participation in global summits and the establishment of a global network of AI Safety Institutes to enhance technical safety measures [7]. The National Institute of Standards and Technology (NIST) has released the “Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile” and continues to develop guidelines for safe and trustworthy AI [1], incorporating PETs [1]. The Department of Homeland Security (DHS) is set to report on actions taken regarding global AI standards and critical infrastructure within 90 days. The National Telecommunications and Information Administration (NTIA) has been tasked with evaluating the risks and benefits of open model weights [3] [4], with a report suggesting that new restrictions should not be imposed without substantial evidence of risk [3] [4].

Overall, while significant progress has been made [1] [5] [8], many issues remain unresolved [8], and the potential for a shift in policy direction under a new administration could impact ongoing AI initiatives [8]. The global landscape of AI development continues to evolve independently of US regulatory frameworks [8], with the administration emphasizing the importance of AI education and training [2], investing in programs to cultivate a skilled workforce [2], and engaging in international discussions to establish global norms for AI governance [2]. The long-term effectiveness of the EO will hinge on policymakers’ ability to adapt to the rapid evolution of AI while fostering both innovation and public trust [6], ensuring that safety [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] measures keep pace with technological advancements.

Conclusion

The Executive Order on AI has led to substantial advancements in AI governance, safety, and ethical standards across various sectors. However, unresolved issues and potential policy shifts under a new administration could influence the trajectory of these initiatives. The ongoing evolution of AI on a global scale necessitates continuous adaptation and collaboration to maintain innovation and public trust while ensuring safety and ethical standards are upheld.

References

[1] https://www.enveil.com/the-ai-executive-order-one-year-reflections/
[2] https://www.whitehouse.gov/briefing-room/statements-releases/2024/10/30/fact-sheet-key-ai-accomplishments-in-the-year-since-the-biden-harris-administrations-landmark-executive-order/
[3] https://cdt.org/press/cdt-marks-one-year-anniversary-of-president-bidens-executive-order-on-artificial-intelligence/
[4] https://cdt.org/insights/press-release-cdt-marks-one-year-anniversary-of-president-bidens-executive-order-on-artificial-intelligence/
[5] https://www.marion-square.com/blog/ai-eo-update
[6] https://www.brookings.edu/articles/one-year-later-how-has-the-white-house-ai-executive-order-delivered-on-its-promises/
[7] https://www.commerce.gov/news/press-releases/2024/10/commerce-marks-one-year-anniversary-historic-biden-harris
[8] https://www.jdsupra.com/legalnews/federal-agencies-take-action-on-ai-one-5167710/
[9] https://www.mintz.com/insights-center/viewpoints/54731/2024-11-07-federal-agencies-take-action-ai-one-year-after-ai
[10] https://www.meritalk.com/articles/agencies-full-steam-ahead-one-year-after-bidens-ai-eo/