The initiative stems from a June 2 executive order that urged advanced AI developers to grant the government early access to their capabilities to address potential vulnerabilities.
The White House on Tuesday launched an AI-backed clearinghouse designed to consolidate software vulnerability findings from government and industry, prioritize the most consequential flaws and coordinate remediation across critical infrastructure.
The initiative, dubbed Gold Eagle, has already begun processing vulnerability reports, the administration said.
Gold Eagle was established under a June 2 executive order that, in part, directed the federal government to promote the development and secure use of advanced AI systems. The initiative brings together the White House, Cybersecurity and Infrastructure Security Agency and the departments of Treasury and Defense with unnamed open-source software organizations and critical infrastructure providers.
The White House press release announcing the initiative says that those partners had developed a system to “receive and patch” vulnerabilities, though it largely describes Gold Eagle as a coordination mechanism. It does not indicate that the initiative would compel companies to address vulnerabilities directly.
The administration also did not specify which agency would oversee its daily operations, how sensitive vulnerability information would be protected or how the initiative would interact with CISA’s existing vulnerability-disclosure and remediation programs. Gold Eagle joins several existing federal vulnerability programs, including CISA’s disclosure program and exploited-vulnerability catalog, the CVE system and NIST’s National Vulnerability Database.
The announcement did not disclose how many findings Gold Eagle has processed, which companies are participating or whether any vulnerabilities have resulted in completed patches.
But Anthropic is likely to be among the initiative’s private-sector participants, given its past commitments to support vulnerability disclosure efforts. The AI company said last month that it would provide federal officials with advance access to its threat-intelligence reports and participate in the interagency vulnerability clearinghouse created by Trump’s June executive order.
“When significant jailbreaks or misuse patterns are identified, we will quickly investigate, triage, and notify appropriate government counterparts,” the company wrote in a June 30 blog post after a recent export control spat with the White House. “We will also provide government partners with our threat intelligence reporting in advance of publication and participate in the interagency cybersecurity vulnerability clearinghouse established under Sec. 2(d) of the June 2 Executive Order.”
Anthropic did not respond to a request for comment by the time of publication.
Increased coordination between industry and government to find and fix vulnerabilities with the help of AI tools follows the spring debut of Anthropic’s Mythos, a powerful cyber-focused AI model. Initially only available to select private sector partners via Anthropic’s Project Glasswing, multiple companies and some federal agencies have been granted access to Mythos for various testing efforts. Releases of cyber-AI models from other major providers have also followed.
Agencies have also recently been placed under more intense pressure to close cyber openings quickly. CISA recently revamped its remediation timeline guidance, ranging from three days for the highest-risk flaws to 60 days for lower-priority issues.