EFF Testifies on Protecting Constitutional Rights from Government AI

The Electronic Frontier Foundation (EFF) brought its expertise to Capitol Hill this week, with Senior Policy Analyst Dr. Matthew Guariglia testifying before the House Homeland Security Subcommittee on Cybersecurity and Infrastructure Protection. Dr. Guariglia delivered a sharp, legally grounded critique of unchecked government AI deployment, centering on two fundamental problems: AI-powered mass surveillance dramatically amplifies existing constitutional rights violations, and pervasive government secrecy around AI systems makes it impossible for the public—and even lawmakers—to audit when these models fail. The testimony cuts through the hype around government AI by focusing on concrete harms rather than theoretical risks, calling for binding safeguards that prevent agencies from using opaque or biased systems to make decisions about citizens' lives.

The EFF's intervention comes at a pivotal moment as Congress drafts legislation shaping how federal agencies can purchase, build, and deploy AI tools. Without transparency requirements, Dr. Guariglia warned, agencies could deploy error-prone or discriminatory models for tasks like immigration enforcement, criminal investigations, and benefits determinations—without any mechanism for affected individuals to challenge or even learn about machine-driven decisions. The EFF is advocating for clear statutory safeguards: rigorous pre-deployment testing, public disclosure of AI use cases, human review of consequential decisions, and protection against mass surveillance dragnets. As lawmakers rush to regulate AI, the message is clear: technology must serve all people, not just the powerful, and rights cannot be sacrificed for efficiency.