White House launches 'Gold Eagle' to coordinate AI-driven cyber defence across critical infrastructure

The White House has unveiled a new initiative designed to strengthen the United States' cyber defences as artificial intelligence rapidly transforms the way software vulnerabilities are discovered and exploited. Named Gold Eagle, the platform will serve as a central coordination hub where government agencies, AI developers and operators of critical infrastructure can share information and jointly respond to emerging cyber risks.

The project reflects growing concern within Washington that advances in AI are dramatically changing the cybersecurity landscape. New-generation AI models are increasingly capable of identifying weaknesses in software at a speed and scale that were previously impossible, creating opportunities for both defenders and malicious actors.

Gold Eagle is intended to ensure that organisations responsible for protecting essential services—from financial institutions to utilities—can coordinate their efforts rather than working in isolation when new vulnerabilities emerge.

More from Tech

A shared platform for faster cyber response

The initiative is being developed jointly by the US Treasury Department, the Department of Homeland Security and the Pentagon. It will bring together government experts, cybersecurity firms, AI companies and selected operators of critical infrastructure to exchange threat intelligence and coordinate technical responses.

Speaking during a media briefing on Tuesday, a senior White House official said the capabilities of modern AI systems have significantly accelerated vulnerability discovery.

"These new capabilities make vulnerability discovery at a scale … that we have not seen before," the official said.

Rather than allowing multiple organisations to independently investigate the same security flaws, Gold Eagle aims to streamline the process by verifying reported vulnerabilities, assigning priorities and coordinating remediation efforts.

According to the White House, the platform will help "deconflict and make sure resources are not being wasted, fixing or scanning for the same vulnerabilities, that those vulnerabilities are validated," while enabling teams of government and industry engineers to jointly assess and resolve the most pressing security issues.

Officials did not identify the companies participating in the programme but said the effort includes American critical infrastructure operators and organisations involved in open-source software development.

Growing focus on AI security

The announcement comes as governments and technology companies grapple with the security implications of increasingly powerful AI models. Several leading AI developers have already delayed broader releases of advanced models to allow trusted partners time to patch vulnerabilities that the systems were capable of uncovering before making the technology more widely available.

Open-source software is expected to be a key focus of the new initiative. Because its source code is publicly available, it underpins a significant portion of the world's digital infrastructure. However, many widely used open-source projects are maintained by small teams or volunteers with limited resources, making it difficult to identify and fix security weaknesses quickly.

The importance of securing open-source software became evident in 2021, when a major vulnerability exposed hundreds of millions of devices worldwide to potential cyberattacks and prompted an urgent response from the Biden administration.

Gold Eagle also fulfils a requirement outlined in an executive order signed by US President Donald Trump in June, which directed the government to establish a formal mechanism for coordinating AI-related cybersecurity efforts.

The same executive order also calls for a framework requiring AI developers to submit their most advanced models to the federal government for review up to 30 days before those models are shared with other trusted partners. That review system is expected to be in place by early August, although the administration has yet to publish its operational details.

Even before the framework is finalised, the White House has intervened in the rollout of advanced AI systems through other measures. Previous actions have included export restrictions affecting Anthropic, which were later lifted, as well as requests that OpenAI delay the broader release of one of its latest models.

Those interventions have fuelled debate within the AI industry, where companies have increasingly argued that a clearer and more predictable regulatory framework is needed. With Gold Eagle, the administration is attempting to build a more structured approach to AI-enabled cybersecurity at a time when both the opportunities and risks associated with the technology are expanding rapidly.