EFF Testifies Before Congress on Protecting Rights Against Government AI

• EFF Senior Policy Analyst Dr. Matthew Guariglia testified before the House Homeland Security Subcommittee on the risks of government AI adoption.
• Key warnings include the use of AI for mass surveillance, potential civil liberty violations, and the lack of transparency in proprietary AI models.
• The testimony emphasized that secrecy regarding government AI usage prevents public oversight and accountability for critical errors.

Testimony Highlights

• Dr. Guariglia argued that the core issue is not controlling AI technology itself, but rather establishing oversight for the government agencies that deploy it.
• AI-driven errors, such as incorrect legal citations or improperly trained recruits at the Department of Homeland Security (DHS), demonstrate the risks of unchecked implementation.
• Proprietary "black box" technology combined with government classification limits the ability of lawmakers and the public to evaluate potential harms to critical infrastructure and individual rights.

Call to Action

• Government adoption of AI must be paired with strict constitutional safeguards.
• Greater transparency is required to prevent secrecy from shielding the government from accountability when AI models fail.