The AI Cyber Fix Washington Can’t Afford To Botch

From the export controls on Anthropic’s latest models to President Trump’s executive order on AI security, the White House has now shifted its focus to AI-enabled cyber security risks.

On July 14, the administration announced Gold Eagle, a government-industry vulnerability coordination initiative. Gold Eagle focuses on finding and fixing cybersecurity vulnerabilities that have become much more visible to potential attackers thanks to AI-driven probes that find weak points far faster than any human analyst could. These include system weaknesses that attackers can exploit to steal information, disrupt operations, or gain unauthorized access. If implemented effectively, Gold Eagle could significantly improve the federal government’s ability to respond to AI-enabled cyber threats at scale.

AI Demands Faster Cybersecurity Coordination

Historically, government agencies have relied on security contractors, vulnerability disclosure programs, and bug bounties to identify flaws and provide time for remediation. As AI accelerates vulnerability discovery, however, the volume may soon exceed what agencies can patch quickly, forcing them to accept that they cannot fix everything at once and only triage the most severe flaws.

Project Glasswing, a coalition launched by Anthropic and major technology companies, demonstrates the scale of the challenge facing the federal government. Traditionally, cybersecurity teams have been constrained by how quickly human researchers could identify vulnerabilities. AI has changed that dynamic. Within its first month, Anthropic and its partners used AI to identify more than 10,000 high- or critical-severity vulnerabilities across important digital systems.

The effort revealed a new cybersecurity bottleneck: organizations discovering vulnerabilities faster than they can verify, disclose, and patch them. For the federal government, cross-government prioritization will be the only way to keep pace.

Jointly operated by the departments of Treasury, Homeland Security, and Defense, Gold Eagle brings together AI developers, cybersecurity firms, open-source software partners, and critical infrastructure operators to coordinate the identification and remediation of vulnerabilities. Its goal is to reduce duplicative vulnerability scanning and provide defenders across government and industry with prioritized, actionable information.

Washington Must Safeguard Gold Eagle Before Adversaries Test It

Gold Eagle’s success depends on creating the legal framework necessary for private sector participation.

The initiative relies on companies voluntarily sharing newly discovered vulnerabilities, model behavior, and other proprietary technical information. Industry partners will hesitate to disclose that information if doing so creates liability, antitrust concerns, or reputational risk.

Congress confronted similar barriers when it enacted the Cybersecurity Information Sharing Act of 2015, which provided liability protections to encourage voluntary cyber threat sharing. Gold Eagle will require comparable statutory safe harbors and clear legal authority for sharing AI-related vulnerabilities while protecting sensitive business information.

Washington must also build the technical workforce needed to turn information into action. Determining which vulnerabilities pose the greatest risk requires experienced AI and cybersecurity professionals who remain in short supply across government. Creating flexible pathways to leverage private-sector experts to support vulnerability assessment and incident response would help ensure Gold Eagle can keep pace with the speed of AI-enabled threats.

Finally, Gold Eagle itself must be secured. By aggregating sensitive vulnerability information from AI developers, cybersecurity firms, and government agencies, the initiative will become a high-value cyber target. The administration should build the platform with zero-trust architecture, strict access controls, and continuous monitoring from the onset. These protections will be essential to ensuring that Gold Eagle strengthens the nation’s cyber defenses rather than creating a new point of systemic risk.

Leah Siskind is director of impact and an AI research fellow for the Center on Cyber and Technology Innovation (CCTI) at the Foundation for Defense of Democracies (FDD). Nidhi Ummettala is a CCTI intern. For more analysis from Leah and FDD, please subscribe HERE. Follow FDD on X @FDD and @FDD_CCTI. FDD is a Washington, DC-based, nonpartisan research institute focusing on national security and foreign policy.