As the use of artificial intelligence in the federal government grows, several agencies have fallen behind with their deadlines and management.
The Government Accountability Office’s (GAO) latest report states that “despite certain federal agencies taking initial steps to comply with guidance and statutory requirement, key efforts to improve management of AI are still not complete.”
The report includes dozens of recommendations for 19 agencies on how to implement AI requirements that are mandated by the federal government. These recommendations include developing plans for agencies to apply the technology, using inventories that include the required information, and issuing guidance on AI use.
The report revealed that 20 out of 23 agencies had reported around 1,200 challenges or opportunities, either current or planned, that could be resolved or assisted by AI. Some examples include analyzing images captured by drones and cameras at borders. Over 200 AI applications are already in use.
NASA, the Departments of Commerce, Energy, Health and Human Services, and State, were the top reported AI users in FY 2022.
Five agencies provided complete information on their AI use cases. The other fifteen, however, had incomplete or inaccurate data. They lacked required data elements such as the life cycle of artificial intelligence or whether or not the AI case could be released. Other AI-related uses were found to not be AI.
The report stated that “without accurate inventories, government management of AI use will be hindered” by incomplete and inaccurate information.
The federal government and Congress are grappling with the implications, benefits, and threats of the technology.
In the last few months, there has been a rush of legislation introduced in Congress to regulate AI. There have also been bipartisan briefings with tech experts and leaders.
The bill, introduced this week by Rep. Randy Feenstra (R-Iowa), would enforce standards in AI programs for the agricultural sector, protecting food, fuel, and other necessities.
The State Department announced last month, that 45 countries have partnered with the U.S. to implement a declaration regarding “responsible use of AI” by military forces.
