Kelly Secures AI Oversight Provisions in Annual Intelligence Bill Passed Out of Committee

Bill also includes Kelly provision targeting uncleared access to classified intelligence

Arizona Senator Mark Kelly, a member of the Senate Select Committee on Intelligence (SSCI), announced that he secured key national security priorities in the annual intelligence bill, the Intelligence Authorization Act for Fiscal Year 2027 (IAA), which passed out of the Committee with strong bipartisan support. The bill includes measures led by Kelly to bring transparency and accountability to the Intelligence Community's use of artificial intelligence (AI), including in decisions with potentially lethal consequences, and improve oversight of access to classified intelligence information across administrations.

Kelly secured five amendments focused on providing Congress transparency into how the intelligence community is deploying AI, establishing safeguards around AI's role in intelligence and in military targeting, and directing new research to get ahead of potential risks AI poses before adversaries can exploit them.

"As AI takes a bigger role in our intelligence and military operations, Congress has a responsibility to understand how it's being used and build in safeguards for the American people before problems arise," said Kelly. "We can continue to lead on AI and be smart about it. That's why I worked to make sure we're using AI responsibly, that Congress stays in the loop, and that we protect our national security from foreign adversaries who are investing heavily into AI."

This year's IAA includes the following Kelly-authored amendments that would:

• Require a report to Congress on newly considered uses of AI, including risk mitigations, future testing, and what risks and advantages the proposed use presents.
• Mandate the Director of National Intelligence (DNI) review how AI is currently being used in targeting workflows and assess the adequacy of existing safety policies.
• Ensure all outputs from an AI system be clearly marked to indicate that AI was used in a way visible to all users, including military personnel, prompting additional validation where appropriate.
• Direct Intelligence Advanced Research Projects Activity (IARPA) studies on AI-on-AI attacks and how AI could contribute to inadvertent escalation with foreign adversaries.
• Establish recurring reporting to Congress whenever the executive branch grants non-clearance holders access to classified information outside established processes.

The bill also focuses on the following areas:

• Significantly reform the IC's authorities to focus on foreign intelligence, international terrorism, and foreign threats to the homeland.
• Provide for additional resources to recruit intelligence officers and expand foreign intelligence collection against U.S. adversaries.
• Prohibit the IC from using or procuring products and services made in Communist China.
• Establish a task force to identify indications and warning of Chinese military aggression regarding Taiwan.
• Enhance oversight and governance of use of artificial intelligence by the IC, including in adoption of advanced AI capabilities and use of AI in lethal targeting.
• Establish opportunities for pre-deployment testing of AI models, helping U.S. firms protect against potential misuse of their models for foreign hacking or weapons proliferation activity.
• Amend the Spectrum Relocation Fund authorization to clarify eligibility for Title 50 agencies that utilize spectrum and whose usage could be impacted by future reallocation decisions.
• Prohibit the IC from acquiring or operating autonomous ground systems from China, Russia, Iran, or North Korea.
• Ensure intelligence support to U.S. allies, including Ukraine, Israel, and Taiwan.
• Protect whistleblowers' right to bring their complaints to Congress in a timely fashion.
• Improve anonymized biological data and intelligence within the IC.
• Establish a Counterintelligence Office at the Department of Commerce.
• Enhance U.S. person privacy by codifying IC policy guidance requiring IC elements to develop procedures for unmasking requests of U.S. person identities.
• Prohibit IC employees and contractors with security clearances from using nonpublic information to place bets on prediction markets.
• Require the IC to report on hostile foreign cyber actors, including foreign scam centers.
• Establish an Office of Counterintelligence within the Department of Treasury's Office of Intelligence and Analysis.
• Create a fellowship named after Senator Ben Sasse for IC employees to gain experience at technology companies and bring this expertise back to their home agencies.
• Mandate security procedures for Federal Reserve data and information.
• Require the Committee on Foreign Investment in the United States to review real estate transactions with foreign persons near IC facilities.
• Extend authorities for protecting Central Intelligence Agency facilities from unmanned aircraft systems.