Artificial intelligence is fundamentally reshaping the cybersecurity landscape — simultaneously arming defenders and handing adversaries powerful new attack tools — technology experts and lawmakers warned during a US Congressional hearing on 7 June 2025. The session, convened by the House Homeland Security Subcommittee on Cybersecurity and Infrastructure Protection, underscored mounting anxiety in Washington over AI-enabled threats to critical infrastructure and the intensifying technology rivalry with China.

The Hearing's Central Warning

The hearing followed President Donald Trump signing an executive order directing federal agencies to develop a framework for evaluating advanced AI cyber capabilities and expanding frontier AI model access for government and critical infrastructure operators. Subcommittee Chairman Andy Ogles set the tone early. 'These models are already reshaping the threat landscape, and the federal government cannot be the last to understand what they can do,' he said.

AI Already Weaponised by Criminals

Sandra Joyce, Vice President of Google Threat Intelligence Group, told the subcommittee that AI has moved beyond theoretical threat status. 'Those concerns were validated recently when we discovered evidence for the first time that AI was used to develop a zero-day exploit by cyber criminals,' she testified. Joyce warned that attackers are now leveraging AI to identify vulnerabilities and traverse networks at speeds that outpace conventional security response. 'Threat actors are able to move rapidly before and after gaining access to a network using AI. They can take advantage of vulnerabilities faster than we can patch, and they can move rapidly through networks using autonomous agents,' she said.

Jack Cable, Chief Executive Officer of Corridor Security and a former adviser at the Cybersecurity and Infrastructure Security Agency (CISA), reinforced that assessment. 'The central challenge is not that AI creates new categories of vulnerabilities. It's that AI dramatically increases the speed and scale at which vulnerabilities can be introduced, found, and exploited,' he said. Cable argued that the focus must shift to preventing vulnerabilities before software is deployed, rather than relying solely on post-discovery remediation.

The Double-Edged Nature of Frontier AI

Chris Meserole, Executive Director of the Frontier Model Forum, highlighted the dual-use dilemma at the heart of advanced AI systems. 'An agent that finds zero-day vulnerabilities can protect us in the hands of a defender but expose us in the hands of an attacker,' he said. Meserole raised particular concern over a technique called 'adversarial distillation,' through which foreign actors could replicate the capabilities of cutting-edge AI models while stripping away their safety guardrails. 'Foreign actors can use distillation to accelerate their own AI development and leverage the capabilities they gain against US critical infrastructure,' he warned.

China as the Defining Rival

China emerged as a recurring flashpoint throughout the hearing. Witnesses cautioned that low-cost Chinese AI systems could achieve widespread adoption across software development, cloud computing, and critical infrastructure if the United States fails to maintain its competitive edge. Joyce was blunt: 'I don't think there's a prize for second place in the AI race, nor is there one in the quantum race.' She noted that Chinese cyber groups have already demonstrated the ability to infiltrate critical infrastructure networks, framing AI leadership as a direct national security imperative. This comes amid broader US-China technology tensions, including export controls on advanced semiconductors and restrictions on Chinese software platforms.

Privacy and Civil Liberties Concerns

Democratic lawmakers raised a parallel set of concerns around domestic surveillance. Matthew Guariglia, a senior policy analyst at the Electronic Frontier Foundation, cautioned that AI-powered tools could dramatically expand government surveillance capabilities unless accompanied by stronger legal safeguards and transparency requirements. The warning reflects a growing tension in Washington between national security imperatives and civil liberties protections — a debate that is likely to intensify as AI capabilities advance. With the Trump administration's executive order now in motion, the hearing signals that Congress is moving toward a more structured legislative response to AI-driven cyber risk.