Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Cybersecurity topic
No spam. Unsubscribe anytime.
Witnesses urge pre‑deployment testing and stronger telecom cyber defenses after AI‑enabled incidents
Summary
Academic and industry witnesses recommended mandatory network security requirements, predeployment testing of advanced models, and staffing boosts for federal cyber agencies after citing recent AI‑enabled incidents and older breaches like Salt Typhoon.
Asad Ramzanali, director of AI and technology policy at Vanderbilt's Policy Accelerator, and other witnesses told the committee that AI both increases cyber risk and offers defensive tools—so the federal government must act on two fronts: hardening critical infrastructure and requiring pre‑deployment testing of powerful models.
Ramzanali summarized three recommendations: "1st, require telecommunications providers to secure their networks... 2nd, mandate predeployment testing of leading AI models... 3rd, this committee should open investigation into the OpenAI Hugging Face incident." He also recalled the Salt Typhoon breach of multiple U.S. providers as evidence that telecommunications networks were already vulnerable.
Bob Everson and others emphasized operational steps companies should take: patching, identifying unpatchable products, multifactor authentication, and a zero‑trust architecture. Panelists also urged Congress to reauthorize CISA 2015 to preserve real‑time threat‑sharing authorities.
