Key policy objectives

The NAII plans to establish cooperation between all US Departments and Agencies, together with academia, industry, non-profits, and civil society organisations to achieve its goals via 6 key strategic pillars:

• Innovation
  ‍To strengthen and leverage the US’s existing R&D ecosystem by combining government, academia, and industry, with a focus on innovative public-private partnerships aimed at accelerating AI discoveries.
  
• Advancing trustworthy AI
  ‍A multifaceted approach that includes R&D into key technical challenges, and the development of metrics and standards to evaluate AI trustworthiness. This includes the development of AI technical standards, and governance approaches for the use of AI in the public and private sectors.
  
• Education and training
  ‍In order to prepare the American workforce for pervasive AI systems in the public and private sectors, the NAII identifies the need for training and development for its citizens according to the goals of the National Science and Technology Council’s Committee on STEM Education (CoSTEM).
  
• Infrastructure
  ‍In the hope to grow the American AI R&D community, the National AI Research Resource (NAIRR) plans to share computing and data infrastructure designed to provide AI researchers access to computation resources and high-quality data, along with appropriate educational tools and user support.
  
• Applications
  ‍To facilitate the continuous R&D and application of AI in all government sectors, including agriculture, finance, healthcare, transport, science, crisis response, national security and defence, and more.
  
• International cooperation 
  To promote a global accord through multi-stakeholder initiatives and bilateral partnerships that supports AI R&D, opens markets for US AI industries, and promotes the development of trustworthy global AI technology.
  

Policy outline

The White House Office of Science and Technology Policy has put forward 10 principles to consider when formulating regulatory and non-regulatory approaches to the development and use of AI in the USA:

• Establish public trust in AI
• Encourage public participation and public awareness of AI standards and technology
• Apply high standards of scientific integrity and information quality to AI and AI decisions
• Use transparent risk assessment and risk management approaches in a cross-disciplinary manner
• Assess full societal costs, benefits, and other externalities in considering the development and deployment of AI
• Pursue performance-based and flexible approaches so as to adapt to the rapidly changing nature of AI
• Evaluate issues of fairness and non-discrimination in AI application
• Determine appropriate levels of transparency and disclosure to increase public trust
• Maintain controls to ensure confidentiality, integrity and availability of AI data such that the AI developed is safe and secure
• Encourage inter-agency coordination to help ensure the consistency and predictability of AI policies

AI risk assessment

The USA’s proposed approach to AI risk assessment can be classified into 3 categories: assessment, independence and accountability, and continuous review.

Assessment

Businesses will be required to conduct algorithm impact assessments to identify the associated risks. The primary objective is to mandate clear descriptions of each risk generated by a particular AI system and then clearly indicate how these risks are being minimised and resolved.

The US State of Virginia’s Consumer Data Protection Act (2021) already requires assessments for some high-risk algorithms. In the EU, the GDPR currently requires similar impact assessments for high-risk processing of personal data.

Independence and accountability

Impact assessments must be conducted by independent third-party players to ensure complete transparency and accountability. Independent reviews and assessments is a cornerstone of transparent AI regulation.

Continuous review

As AI evolves at a rapid rate, continuous review and impact assessment of AI systems is crucial in remaining on the edge of compliance at the rate of innovation. Thus, no single assessment will suffice.