Landscaping AI and Security at The House
Yesterday, testimony was heard on “The AI Security Landscape: How AI is Reshaping Cybersecurity and Critical Infrastructure Resilience” by the House Homeland Security Committee Subcommittee on Cybersecurity and Infrastructure Protection. The four expert witnesses each testified to a theme giving shape to risks and opportunities of the security applications for large language models. Sandra Joyce, Vice President of Google Threat Intelligence, called for greater regulatory mechanisms. Chris Meserole, Executive Director of the Frontier Model Forum rose concerns over distillation. Jack Cable, CEO and Cofounder of Corridor, brought attention to the lack of open weight models originating in the U.S. And Matthew Guariglia, Senior Policy Analyst with the Electronic Frontier Foundation, inveighed against the heightened threat of domestic surveillance and vulnerability stockpiling.
Before the session came to a close, Ranking Member Delia Ramirez asked a final question that went largely unanswered: who should be held accountable when AI goes rogue? Perhaps that will be the subject of the next hearing.
If you found your way to this post, I encourage you to watch the whole thing and form your own opinions about how AI could reshape security. This is a political project, which means we are all in this together, so more voices heard means a better chance at carving out a path forward that works for us all.