The Office of Inspector General of the Intelligence Community (ICIG) raised concerns over oversight of artificial intelligence (AI) initiatives and how inspectors general can keep up with the pace of technology in the ICIG semiannual report to Congress, publicly released today.
The report notes that IT is the first listed programmatic objective for the intelligence community (IC), with both cybersecurity and modernization at the forefront.
During the first half of fiscal year 2019, ICIG found weaknesses in management of privileged users at the Office of the Director of National Intelligence (ODNI), that the Cyber Threat Intelligence Integration Center had not fully consolidate with the National Intelligence Manager for Cyber, and started reviews on supply chain risks in acquisition and the implementation of the Cybersecurity Information Sharing Act.
With AI advancing at a fast pace, the ICIG report also raised some concerns that oversight may not be able to keep up.
“Investment asymmetry between mission performance and intelligence oversight in AI efforts could lead to an accountability deficit,” the report warns.
Despite the quick advance of technology, the ICIG report affirmed the idea that oversight is needed and cannot be omitted as priorities on AI are set.
“Reassuring statements that the IC is currently using AI technologies – and will use AI technologies in the future – in ways consistent with the rule of law and American values will not be sufficient. The IC will need to validate those statements for the American people in understandable, timely, objective, and transparent ways,” the report said.
The report noted that while AI is a national priority, as the President’s Executive Order on AI established, the various efforts around AI put the oversight community into new territory. With multiple large projects underway, auditors are working to keep pace, citing the IC’s AIM Initiative, the Department of Defense’s Joint Artificial Intelligence Center (JAIC), and the AI Next campaign from the Defense Advanced Research Projects Agency (DARPA).
“These multiple, costly, and complex efforts will pose profound challenges for intelligence oversight authorities,” the report said. “The ICIG is taking action to build general awareness and common understanding among intelligence oversight authorities.”
Among the activities, the ICIG met with the head of the AIM initiative, focused on the difference between AI and other automated technologies, met with other inspectors general to discuss the issue, and talked with IC officials to gain insights on how oversight can improve when investigating AI.
Future efforts for the ICIG include establishing a community of interest, evaluating investments in AI oversight, and engaging with the Council of Inspectors General on Integrity and Efficiency.