The Government Accountability Office found that some government agencies that presented use cases for artificial intelligence have not met key requirements for implementing the technology.
The government watchdog on Tuesday released a new report presenting 35 recommendations to enable agencies to properly implement and manage AI despite its risks.
The study was conducted among 23 government departments, excluding the Department of Defense, which was subjected to a different AI readiness evaluation. Twenty of them had approximately 1,200 existing and planned AI use cases, while three agencies said they could not pinpoint areas that may be solved by the technology.
NASA submitted more than 300 AI use cases, followed by the Department of Commerce with 285. Requirement implementation has led to the establishment of standards and facilities such as the General Services Administration’s AI Center of Excellence.
However, the report also showed important actions pending in some agencies. For instance, the Office of Management and Budget has not created guidelines for AI use and acquisition. The Office of Science and Technology Policy also lacked specific designation of agencies expected to meet specific AI requirements.
Among GAO’s recommendations were aligning data with guidance when it comes to AI use case inventories and complying with federal law and policy such as developing methodology for annual inventory updates.
The Potomac Officers Club will hold its fifth annual Artificial Intelligence Summit on March 21, 2024. Join federal officials and industry experts as they discuss AI advancements, risks and challenges. Register now!
