On October 4, 2022, the White House Office of Science and Technology Policy published The Blueprint for an AI Bill of Rights: Making Automated Systems Work for the American People (the “AI Blueprint”). The AI Blueprint outlines non-binding guidelines for the development and deployment of automated systems and is the culmination of a year-long process of public engagement and deliberation.
- The AI Bill of Rights is a set of five principles: (1) Safe and Effective Systems, (2) Algorithmic Discrimination Protections, (3) Data Privacy, (4) Notice and Explanation, and (5) Human Alternatives, Consideration and Fallback. Taken together, the principles require companies to consider and evaluate how AI may negatively impact individual rights, opportunities, and access to critical resources. Notably, the AI Blueprint also directs companies to provide clear and simple notice regarding their use of AI, and to allow users to opt-out of an automated process (e., by providing them with the opportunity to engage with a human).
- Data privacy is at the center of the framework. The AI Blueprint notes that “[d]ata privacy is a foundational and cross-cutting principle required for achieving all others in this framework.” It further urges companies to make privacy the default setting in automated systems and to work towards data minimization. Interestingly, the AI Blueprint highlights the lack of a comprehensive federal privacy framework – a recently hot topic with the advancement of the American Data Privacy and Protection Act – stressing that such a framework would afford Americans greater knowledge and control over the impact of automated systems on their data.
- The framework highlights areas where the use of AI is of particular concern, including health, employment, housing, personal finance, education, and policing. The publication therefore builds on the recent trend of government attention to private sector surveillance in sensitive domains.
- Public reporting is an element of each principle. The AI Blueprint calls for reporting under each of the five principles, urging companies to not only take steps to develop and deploy AI within the outlined guidelines, but to document those steps and report progress to the public.
While the 73-page AI Blueprint certainly enumerates a strong position on the use of automated systems, it is not an enforceable document and companies must opt to comply with it. Moreover, entities that choose to meet some or all of its principles do not automatically comply with other current and upcoming AI statutory and regulatory requirements. Nevertheless, the AI Blueprint provides insight into the kinds of issues that may be prioritized by the current administration moving forward and demonstrates the continuing depth of governmental focus on this topic.
This post is as of the posting date stated above. Sidley Austin LLP assumes no duty to update this post or post about any subsequent developments having a bearing on this post.