Key Point: The Biden Administration is creating a new Federal agency focused on the safety and security of Artificial Intelligence systems.
Two days after issuing an Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence (AI), President Biden directed the Department of Commerce to establish the U.S. Artificial Intelligence Safety Institute (USAISI), which will lead the government's efforts on AI safety and trust, particularly for the evaluation of the most advanced AI models.
What will USAISI actually do?
USAISI will facilitate the development of safety, security and testing standards for AI models, and the development of standards for authenticating AI-generated content. Earlier this year, the Administration announced that it would provide $100 million in funding to USAISI over the next five years. For comparison’s sake, the Office of the National Cyber Director was initially stood up with $250,000 in funding pulled from other executive offices.
Led by the National Institute of Standards and Technology (NIST), USAISI will work with partners in academia, industry and government to advance AI safety. Additionally, USAISI will coordinate with the United Kingdom’s AI Safety Institute and similar institutes in ally and partner nations. While USAISI appears to have a civilian/private sector focus, presumably it will coordinate with the National Security Agency’s recently announced AI Security Center.
The NSA’s new organization is charged with developing best practices, evaluation methodology, and risk frameworks to promote the adoption of AI capabilities across the national security enterprise and the defense industrial base. The NSA Director, General Nakasone, recognizes the importance of collaboration and government-wide emphasis on AI technology, stating "AI will be increasingly consequential for national security in diplomatic, technological and economic matters for our country and our allies and partners."
What are the legal implications of a new Federal agency focused on AI systems?
USAISI has the potential to drive important consequences for industry on the legal and compliance fronts. For example, USAISI-developed safety standards could become the baseline used by courts and regulators to determine whether the developers (or users) of an AI system were negligent in the system’s creation (or operation), or whether an AI system was used to infringe on the intellectual property of a third party.
Similarly, USAISI guidelines and standards could influence whether states’ product liability laws will be applied to AI systems, or whether a company’s information security program and incident response plan were sufficient following a data security incident.
On the consumer and employee front, USAISI's research on AI safety and security will likely influence the creation of new legal frameworks that are intended to protect privacy interests and economic interests of individuals.
Conclusion
The establishment of USAISI could be a significant development for this emerging technology sector, particularly as the fields of AI, Operational Technology and Robotics intersect. Private industry should monitor USAISI’s progress for opportunities to partner with the agency and advocate for outcomes that properly balance the interests of all stakeholders.