American Civil Liberties Union

The American Civil Liberties Union (ACLU) is the United States' oldest and largest civil-liberties organisation — a non-profit founded in 1920 in response to the Palmer Raids, with a current membership of over 1.7 million, more than 500 staff attorneys, approximately 2,000 volunteer attorneys, a near-2,000-cases-per-year caseload, and a 50-state network of staffed, autonomous affiliate offices headquartered in New York. In the make-AI-good movement's terms, the ACLU occupies a distinctive position: it is the US national civil-liberties infrastructure whose constitutional-law depth, litigation capacity, and affiliate network have made it the primary vehicle through which US communities have won enforceable democratic oversight of police surveillance technology. The Community Control Over Police Surveillance (CCOPS) model-ordinance campaign — co-launched with the Electronic Frontier Foundation in September 2016 and adopted in 26 jurisdictions protecting nearly 18 million people by 2022, including the first US municipal facial recognition ban (San Francisco, 2019) and the New York City adoption (2020) — is the corpus's clearest US example of a civil-liberties organisation converting constitutional principle into enforceable local democratic governance of AI-powered surveillance.

Founding and structure

The ACLU was founded in 1920 in direct response to the Palmer Raids — the Justice Department's mass arrests of thousands of people "without warrants and without regard to constitutional protections against unlawful search and seizure" across November 1919 and January 1920 — when, as the organisation's own history records, "a small group of people decided to take a stand, and thus was born the American Civil Liberties Union." Roger Baldwin, a civil-liberties activist and pacifist who had been imprisoned as a conscientious objector during World War I, was the ACLU's founding director and shaped its founding posture: that the Bill of Rights belongs to everyone, including the government's most unpopular targets, and that defending constitutional principles requires defending them at the hardest case, not just the convenient one. That posture — universal civil-liberties defence grounded in constitutional litigation and public advocacy — has defined the organisation through more than a century of contested terrain.

The ACLU's structure is a federated model: a national office in New York and 50 state-level affiliate offices, each staffed and autonomous, that gives the organisation simultaneous national litigation capacity and local political reach in every US state. Anthony Romero, the ACLU's Chief Executive Officer, has led the organisation since 2001 — a tenure spanning the post-9/11 expansion of domestic surveillance, the mass surveillance revelations of the Obama era, the Trump-era immigration and civil-rights campaigns, and the current AI-governance decade. Deborah Archer serves as President. The national organisation handles close to 2,000 cases annually through its staff and volunteer attorney network; the affiliate network multiplies that capacity across the full US geography, including the Technology for Liberty programmes at state affiliates like ACLU of Massachusetts that run the city-level campaigns that CCOPS depends on.

Speech, Privacy, and Technology Project

The ACLU's Speech, Privacy, and Technology Project is the operational unit whose work sits most directly in the make-AI-good movement. The project's programme runs across eight areas — consumer privacy, internet privacy, location tracking, medical and genetic privacy, national identification systems, border and checkpoint privacy, surveillance technologies, and workplace privacy — and combines courtroom litigation, legislative advocacy on Capitol Hill, and direct engagement with technology companies. The ACLU's standing diagnosis of its operating environment is that "electronic privacy laws haven't been updated since 1986, predating the modern internet" — a constitutional and statutory framework designed for a pre-digital world in which courts, legislatures, and administrative agencies are making real-time decisions about AI-powered surveillance under law that was never designed to govern such systems.

Jay Stanley, Senior Policy Analyst in the Speech, Privacy, and Technology Project, has authored and co-authored hundreds of influential ACLU reports, white papers, and blog posts on privacy and technology, edits the ACLU Free Future blog, and has shaped the public framing of surveillance-technology risks in the United States — from the 2019 "Dawn of Robot Surveillance" report through ongoing work on facial recognition, AI-generated police reports, and digital identity legislation. At the affiliate level, Kade Crockford, Director of the Technology for Liberty Project at the ACLU of Massachusetts and an MIT Media Lab Director's Fellow, runs the affiliate-level work that produced the Massachusetts CCOPS cluster — Cambridge (2018), Somerville (2019), Boston, Lawrence, and Northampton — through sustained city-level organising with civil-liberties, racial-justice, and community coalitions.

CCOPS and democratic oversight of police surveillance

The ACLU's most structurally significant contribution to the make-AI-good movement's US chapter is the CCOPS model-ordinance campaign. The ACLU California affiliates developed the model local surveillance ordinance beginning in 2014 and ran a statewide California campaign; the first ordinance passed in Santa Clara County in 2016, through a coalition including the EFF, the ACLU of Northern California, CAIR San Francisco Bay Area, and Oakland Privacy. The national campaign launched formally on September 21, 2016, co-led by the ACLU and EFF with a founding coalition that included the Bill of Rights Defense Committee/Defending Dissent Foundation, Demand Progress, Million Hoodies Movement for Justice, NAACP, National Network of Arab American Communities, Restore the Fourth, and South Asian Americans Leading Together.

The CCOPS model ordinance's core demand is that city councils — not police agencies — must pre-approve the acquisition and deployment of any surveillance technology, after a public process including a civil rights and civil liberties impact assessment, community input, and a council finding that the technology's benefits outweigh its costs. The ordinance's eight guiding principles — pre-approval, meaningful community role, transparent decision process, technology-specific approval, civil rights impact assessment, financial-implications consideration, annual public reporting, and no grandfathering of existing systems — were designed to be portable across politically diverse US municipalities without requiring a centrally-resourced campaign in each city. By 2022, 26 US jurisdictions had adopted CCOPS laws protecting nearly 18 million people: San Francisco (2019, the first US municipal ban on facial recognition technology, enacted through the Stop Secret Surveillance Ordinance); New York City (2020, the largest US city to adopt a CCOPS-framework law, via the POST Act requiring the NYPD to publicly disclose its surveillance technologies and policies); and the Massachusetts cluster produced through Kade Crockford's Technology for Liberty programme.

AI surveillance research and litigation

The ACLU has produced foundational documentation of AI surveillance risk in the US context. "The Dawn of Robot Surveillance", published June 13, 2019, analysed how AI and video analytics were transforming surveillance cameras from passive recording devices into active monitoring systems capable of human-action recognition, anomaly detection, physiological measurement, and emotion recognition — with the NYPD and Microsoft already deploying the technology across more than 6,000 cameras. The report's concern was that mass surveillance of this kind — continuous, automated, population-scale — would chill public behaviour, deepen racial and socioeconomic inequalities in policing, concentrate power between government and citizens, and operate with no transparency or algorithmic accountability; its policy recommendation was that policymakers "prohibit mass surveillance applications and establish safeguards against abuse."

On facial recognition specifically, the ACLU's litigation programme has documented a pattern of wrongful arrests caused by facial recognition misidentification — including cases where ACLU clients spent months in jail on the strength of an incorrect facial recognition match before the error was identified. These cases form the empirical spine of the ACLU's argument that facial recognition technology in law enforcement is not merely a civil-liberties concern in principle but an active source of serious harm to identifiable individuals. The ACLU has also documented active defiance of local prohibitions, as in the case of the New Orleans Police continuing to use live facial recognition despite a local ordinance against it — using the technology in violation of the democratic governance framework that CCOPS ordinances represent.

Position in the movement

The ACLU's place in the make-AI-good movement is as the US constitutional infrastructure at scale. Where organisations like the Algorithmic Justice League operate primarily through research and public narrative, the ACLU combines research-and-expose, grassroots legislative organising, and constitutional litigation through a national affiliate network capable of running parallel campaigns in every US state simultaneously. Its theory of change is rooted in the constitutional tradition: that fundamental rights are non-negotiable, that government power over citizens' information and movement must be constrained by law and democratic accountability, and that the courts and the democratic legislature — not administrative discretion or law enforcement self-regulation — are the right venues for establishing those constraints. The CCOPS campaign is the operational expression of that theory: not categorical prohibition but democratic procedure, building the principle of community control over police surveillance technology into enforceable local law, city by city, across a politically diverse national landscape.