{{item.videoDuration}}
{{item.title}}
{{item.text}}
{{item.videoDuration}}
{{item.title}}
{{item.text}}
Background and overview
In order to grasp the potential benefits of, and manage the risks posed by artificial intelligence (‘AI’), on 30 October 2023 the Biden Administration released ‘The Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence1’ (‘the EO’). The EO extends over several sections, covering Federal Agencies and private enterprises across manifold sectors (Figure 1). The EO follows upon the comparatively soft law approach aligned with AI industry development seen in the 2019 Executive order ‘Maintaining American Leadership in Artificial Intelligence2’ and the ‘Blueprint for an AI Bill of Rights3’ released in 2022.
Figure1: Main contents of the EO
| Section name | Main contents |
| Ensuring the Safety and Security of AI Technology |
|
| Promoting Innovation and Competition |
|
| Supporting Workers |
|
| Advancing Equity and Civil Rights |
|
| Protecting Consumers, Patients, Passengers, and Students |
|
| Protecting Privacy |
|
| Advancing Federal Government Use of AI |
|
| Strengthening American Leadership Abroad |
|
| Implementation |
|
Ensuring the Safety and Security of AI Technology
- Requirements for developers of the most powerful AI systems such as foundation models etc. to share red-teaming results with the federal government
- Establishment of standards to ensure the safety and security of AI systems
- Development of new standards relating to, and mitigation of, risks from AI-enabled production of dangerous biological substances
- Establishment of authentication criteria and best practices related to AI output, and protection from fraud, etc.
- Development of AI tools to discover and modify vulnerabilities, and promote improvement of cyber security
- Creation of a memorandum on AI and security, and guarantees on the ethical use of AI by intelligence agencies
The EO directs the NIST to issue guidelines, standards and best practices on AI safety and security. In order to ensure the safety and reliability of AI, companies that develop potential dual-use foundation models and companies that own large-scale computing clusters are to be asked to provide reports on their activities. In addition, in order to deal with the use of US IAAS products by foreign malicious cyber actors, the EO calls for a proposal for additional regulations on the reporting of foreign users of US IaaS products who are opening or maintaining an account with a foreign reseller.
In order to reduce AI cyber security risks for critical infrastructure, each federal agency is to assess the risk related to the use of AI. In addition, the EO directs the issuance of the best practices to manage financial institutions’ AI -specific cyber security risks, the application of the AI Risk Management Framework (AI RMF4), and the establishment of an Artificial Intelligence Safety and Security Board. In addition, the EO calls for evaluating and mitigating the possibility of AI use which presents CBRN (chemistry, biological, radiation, nuclear) risks, through AI biosecurity measures in the nucleic acid synthesis industry.
To improve the identification of synthetic content generated by AI, the EO directs creation of a report to identify existing standards, tools and methods, and to examine the development of further standards and technology. In addition, the EO calls for guidance to be issued, and for the Federal Acquisition Regulations (FAR) to be updated to account for this guidance. Furthermore, the EO directs the soliciting of public opinions to make an assessment of the risks and benefits of dual-use foundation models with widely available weights, and creation of a report on preventing federal data leakage and malicious use.
Promoting Innovation and Competition
- Use of National AI Research Resource to buttress US AI research
- Technical assistance to small scale AI developers and enterprises etc.
- Streamlining of visa procedures, and active pursuit of foreign national AI talent
To attract and maintain talent in AI and other important emerging technological fields within the United States, the EO calls for consideration of a reduction in the burden imposed by visa application and examination procedures, and changes in policy such as clarification and modernisation of immigration pathways for experts in AI and other critical and emerging technologies.
Additionally, directives are made for several actions by federal agencies, such as strengthening of public-private cooperation, support for high-performance computing, clarification of AI and patent-related issues, measures tackling AI-related IP risks, promotion of patient and medical worker welfare, quality improvements in veterans’ healthcare, and strengthening of US climate change resilience.
Supporting Workers
- Mitigation of disadvantages, and maximisation of the benefits, posed by AI to workers
- Evaluation of the impact of AI on the labour market and strengthening of federal government support
In order to support workers and deepen the understanding of AI's effects on the labour market, the EO directs the creation of a report on the labour-market impact of AI, and a report on measures to support the workforce displaced by AI and other technological disruptions. The EO also requires developing and publishing principles and best practices for employers to use to mitigate AI’s potential harms to employees’ well-being and maximise its potential benefits.
Advancing Equity and Civil Rights
- Provision of guidance to federal agencies, leaders etc. to ensure non-discrimination in AI use
- Measures against algorithmic discrimination in the US legal system
- Measures to ensure the fairness of the criminal justice system
Creation of a presidential report and best practices on measures to enhance civil rights and secure fair judicial service is required by the EO. In addition, federal agencies are to issue guidance on illegal discrimination in recruitment arising from AI, the use of automated systems or algorithmic tools used to determine access to housing, and on making the benefits of AI available to the disabled.
Protecting Consumers, Patients, Passengers, and Students
- Promotion of responsible AI use in the area of health and welfare
- Support for educators through the use of AI educational tools etc.
The EO requires independent regulatory agencies to consider the use of their full authorities to protect consumers from fraud, discrimination, privacy infringement and other risks arising from AI. Efforts are to be made to secure responsible AI development and deployment in the healthcare, public-health and human-services sectors, as well as the Departments of Transport and Education. The Federal Communications Committee is also required to consider actions related to the effect of AI on communications networks and consumers.
Protecting Privacy
- Promotion of AI technology development while protecting privacy
- Promotion of privacy enhancing technologies such as encryption tools
- Formulation of guidance on data collection, and (privacy) assessments by the federal government
- Formulation of guidance to evaluate privacy enhancing technologies, including AI
To mitigate the privacy risks entailed by AI, the EO directs creation of guidance on commercially available information with privacy and confidentiality risks. The EO requires guidelines be created to assess the effects of differential-privacy guarantees for AI, and advancement of the development, implementation and application of privacy enhancing technologies (PETs).
Advancing Federal Government Use of AI
- Provision of AI procurement and deployment guidance to federal government organs
- Facilitation of efficient AI products and services acquisition through accelerated contracts etc.
- Acquisition of AI talent for federal government, and provision of training to employees
Each federal agency is to designate a 'Chief Artificial Intelligence Officer,' and take a risk-based approach to the use of generative AI, rather than imposing general bans.
In addition, a 'White House AI Council' will be established to ensure effective formulation, development, communication, involvement with the industry, and timely implementation of AI related policies.
Strengthening American Leadership Abroad
- Strengthening of global cooperation
- Creation of global AI standards
- Engagement on global issues around safe, responsible, rights-affirming AI development
The EO aims for the United States to take the global lead in AI through expanded global cooperation, establishment of a robust global framework for managing the risks and leveraging the benefits associated with AI, and the creation and use of global standards based on the AI RMF.
Conclusion
The EO contains several directives, with force of law, which may affect Japanese enterprises. In particular, regulations related to foreign users of IaaS products may require close vigilance to ensure compliance. Furthermore, activities within the US federal government, such as policy setting and committee/agency establishment, are likely to be incorporated into the policies of other jurisdictions which may be categorised as regulation-takers, and as such those activities will continue to command attention going forwards.
1 Biden Administration, Exec. Order No. 14110, 2023, Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, Viewed 2023/11/30, Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence
2 Trump Administration, Exec. Order No. 13859, 2019, Maintaining American Leadership in Artificial Intelligence, Viewed 2023/11/30, Maintaining American Leadership in Artificial Intelligence
3 Biden Administration, 2022, Blueprint for an AI Bill of Rights, Viewed 2023/11/30, Blueprint for an AI Bill of Rights
4 National Institute of Standards and Technology, 2023, Artificial Intelligence Risk Management Framework (AI RMF 1.0), Viewed 2023/11/30, Artificial Intelligence Risk Management Framework (AI RMF 1.0)
{{filterContent.facetedTitle}}
{{contentList.dataService.numberHits}} {{contentList.dataService.numberHits == 1 ? 'result' : 'results'}}
{{contentList.loadingText}}
{{contentList.loadingText}}
