In a recent development, the American Civil Liberties Union (ACLU) has filed a Freedom of Information Act (FOIA) lawsuit against the National Security Agency (NSA) to obtain records about the NSA's use of artificial intelligence (AI) [efc0bd02]. The ACLU is seeking the release of studies, roadmaps, and reports that shed light on how the NSA is utilizing AI and how it impacts privacy and civil liberties. The lawsuit aims to address the lack of transparency surrounding the NSA's AI tools and their potential to amplify biases and have higher error rates when applied to marginalized communities. By seeking critical information about the NSA's use of AI, the ACLU aims to facilitate a meaningful debate about government surveillance and ensure the protection of civil rights and liberties [efc0bd02].
This FOIA lawsuit comes at a time when the NSA has been rapidly developing and deploying AI, but little is known about its specific applications and effects on civil rights. The government has not released key documents that examine the impact and efficacy of the NSA's AI tools, nor has it disclosed the safeguards in place to protect against potential abuses [efc0bd02].
The ACLU's lawsuit was filed in April 2024 in the District Court (S.D.N.Y.) [efc0bd02].
In another development, the Future of Privacy Forum (FPF) has launched the FPF Center for Artificial Intelligence (AI) to support AI governance and policymaking. The center aims to establish best practices, conduct research, track legislation, and provide information to stakeholders. FPF recognizes the global challenges posed by AI and believes in the importance of international collaboration. The center will include sector-specific working groups and balance local and global perspectives. It will compare different jurisdictional approaches to AI policy and governance and prioritize building expert assessments for evaluating AI implementations. FPF has received funding from the National Science Foundation and the Department of Energy. The center builds on FPF's existing work in AI and its membership in the U.S. AI Safety Institute [de64053e].