President Trump signed an executive order (EO) on Dec. 3 that sets forth guidelines for the use of artificial intelligence (AI) technologies for Federal civilian agencies, and a roadmap to implementing those guidelines that will out-last the current administration set to depart on January 20.
Notably, the order does not apply to the use of AI in defense or national security systems.
According to the White House Office of Science and Technology Policy (OSTP) – which has shouldered much of the administration’s work on AI policy issues – the EO covers four action items to establish: principles for AI use; policies for implementing those principles; creating a catalog of Federal agency AI use cases; and a playbook for increasing AI expertise at the Federal level. Each action item comes with an initial timeline of 180 days or less to begin implementation.
“Artificial intelligence can be an important tool to help modernize government and ensure Federal agencies are effectively and efficiently delivering on their missions on behalf of the American people,” said Michael Kratsios, U.S. Chief Technology Officer and acting under secretary of defense of research and engineering at the Department of Defense.
“The Trump Administration is committed to advancing AI innovation that benefits all Americans and is underpinned by American values,” he said. “This Executive Order will foster public trust in the technology, drive government modernization, and further demonstrate America’s leadership in artificial intelligence.”
The order authorizes the creation of numerous ethics-related principles to govern the use of AI at the Federal civilian agency level. Altogether the order lists nine principles, which boil down to primary themes of not violating Constitutional or civil rights; using only when benefits greatly outweigh assessed risks; using only for specific design reasons; and providing the ability for outcomes to be readily understood.
In addition, the order says AI should be secure from vulnerabilities, the humans in charge of running AI programs should have clear roles, organizations should be transparent about their use of AI, and organizations must be accountable for their uses of AI and regularly test their uses against these principles.
The Office of Management and Budget (OMB) will oversee the creation of implementation guidelines. OMB already has policies on the books that address some of the principles in the EO, but don’t specifically refer to AI. The order gives OMB 180 days to create a game plan to bridge the gap between existing policies and the principles outlined by the order.
The order also authorizes the creation of a database of Federal government use cases for AI, provided they are non-classified and non-sensitive, and asks the Federal CIO Council to create guidelines for the inventory within 60 days. After that, agencies will have to create an inventory of how they are using AI, and share that repository across agencies.
The order’s final action item requires the Presidential Innovation Fellows program to create a path to attract and retain AI industry leaders to work within agencies across government. The program – which gathers tech-trained individuals from the private sector and academia for a year of government service – will have the task of working with other Federal agencies to find the best way to increase interagency institutional knowledge of AI and its applications.