Federal government lagging behind on key AI requirements, watchdog finds
More than a dozen federal agencies are lacking in their implementation of guidelines related to artificial intelligence, a new Government Accountability Office report finds.
A number of federal agencies are lagging behind on deadlines and efforts to better their management of artificial intelligence, as the technology’s footprint is expanding in the federal government.
"Although certain federal agencies have taken initial steps to comply with guidance and statutory requirements, key efforts to strengthen management of AI have missed deadlines and are not yet completed," the Government Accountability Office (GAO) report issued this week says.
The report makes dozens of recommendations to 19 agencies to take additional steps to fully implement requirements on AI that are federally mandated. Those recommendations include the development of plans across agencies on the application of the technology, the use of inventories to include required information and the issuing of guidance on AI usage.
BIPARTISAN LAWMAKERS EYE AI SAFEGUARDS FOR US AGRICULTURE INDUSTRY
The report found that 20 of the 23 agencies reported about 1,200 current or planned challenges or opportunities that could be solved or aided by the use of AI. Examples include analyzing cameras at the border and analyzing images taken by drones. There are over 200 instances in which AI is already being used.
NASA, as well as the Departments of Commerce, Energy, Health and Human Services and State were those with the highest reported AI use cases in FY 2022.
The report found that five agencies had provided comprehensive information for their use cases, but the other 15 had either incomplete or inaccurate data – including lacking required data elements, such as life cycle of the artificial intelligence, or whether the AI case was releasable. There were other instances of AI uses that were later determined not to be AI after all.
"Without accurate inventories, the government's management of its use of AI will be hindered by incomplete and inaccurate data," the report said.
The report comes as the federal government and Congress look to grapple with the implications, advantages and threats posed by the technology.
There has been a flurry of legislation in Congress in the efforts to regulate AI, as well as a number of bipartisan briefings on the matter with key tech leaders and experts.
A bill introduced by Rep. Randy Feenstra, R-Iowa, this week would enforce standards for AI programs in the agricultural sector to protect food, fuel and other necessities.
Meanwhile, the State Deptartment announced last month that 45 foreign governments have now partnered with the U.S. to launch the implementation of a declaration on "responsible military use" of AI.
Fox News' Liz Elkind contributed to this report.