Legal aid, accountability journalism, voting access, public-records tooling, and the algorithmic-accountability research that holds high-stakes government and corporate AI to scrutiny. Where AI most often amplifies state power, this domain documents the counter-instruments.
Algorithmic systems are deciding bail, benefits eligibility, hiring outcomes, and immigration status — while the people subject to those decisions frequently have no path to contest them and no way to know they were made by software. The catalog's job here: surface the research institutions, journalism organisations, and access-to-justice tooling that document the harms and build the counter-infrastructure. The negative-space companion is plain: predictive policing, scoring tools sold to law enforcement, and benefits-fraud algorithms used against claimants do not appear in this catalog.
The institutions doing the work of auditing, naming, and documenting the harms of high-stakes algorithmic systems — with research output that's been load-bearing across journalism, regulation, and litigation.
Joy Buolamwini's organisation. Bias audits of commercial face-analysis systems, advocacy work that fed directly into municipal facial-recognition bans, and the cultural-narrative work (the Coded Bias documentary) that brought algorithmic-accountability into mainstream civic conversation.
An independent research institute on the social implications of AI — labour, accountability, climate, and the political economy of large-scale AI deployment. Annual reports that have functioned as agenda-setting documents for advocates and regulators.
An independent research organisation studying the social implications of data-centric and automated technologies. Long history of qualitative and ethnographic work on how algorithmic systems land in real institutions — courts, schools, child welfare — that quantitative-only research consistently misses.
A Berlin- and Zurich-based nonprofit doing investigative and policy work on automated decision systems in Europe. Strong record of reporting on government ADM systems (welfare, education, policing) and feeding those findings into the European policy process around the AI Act and DSA.
The Distributed AI Research Institute, founded by Timnit Gebru. Independent research on the data, labour, and harm patterns behind dominant AI systems — with a deliberate model of distributed, internationally-rooted research that doesn't run through the typical Bay Area institutional substrate.
The research and advocacy arm of the Mozilla Foundation. Internet Health reporting, the *Privacy Not Included consumer guide, and a steady output of investigative work on dominant platforms — including extensive YouTube algorithm research conducted with volunteer Regrets Reporter contributions.
Accountability journalism that uses AI as a research instrument — not as a content factory. Newsrooms and investigative collectives whose track record on algorithmic-system reporting is the proof of the model.
An investigative collective using open-source intelligence techniques on conflict, organised crime, and human-rights cases. Publishes a public OSINT toolkit and runs training that has effectively become the on-ramp for a generation of independent investigators. The serious version of "verification at scale".
A nonprofit investigative newsroom. Long-running track record on algorithmic-system reporting — the Machine Bias series on COMPAS recidivism scoring is still the canonical entry point into the criminal-justice algorithm conversation. Ongoing data-driven accountability work across the federal government.
A data-driven newsroom focused on tech accountability, now operating as part of CalMatters. Methodology-forward reporting on Citizen Browser, Blacklight, and other tools that turn the data-extraction machinery back on the platforms doing the extracting.
A research agency at Goldsmiths, University of London, using architectural and media analysis to investigate state and corporate violence. Their reconstructions have been used as evidence in courts, UN proceedings, and human-rights tribunals — the working definition of accountability research as forensic practice.
A global digital-rights organisation. Operates the Digital Security Helpline for at-risk users, runs RightsCon, and publishes research on internet shutdowns, spyware, and rights-violating tech deployments — one of the most reliable cross-border early-warning networks for AI-enabled rights abuses.
Making legal systems navigable for non-lawyers. The legal-aid infrastructure, design research, and self-help tooling that exists because the formal legal system has not made itself usable by the people it most affects.
The Legal Design Lab at Stanford Law. Research and open prototypes on access-to-justice problems — eviction, debt collection, civil legal aid — with explicit work on how AI assistance could be deployed in court self-help contexts without producing more harm than it resolves.
A nonprofit that helps low-income Americans file Chapter 7 bankruptcy without a lawyer, free. One of the cleanest examples of legal-aid software meeting a population that the formal pro-bono system can't reach at scale — and one of the better-documented case studies on the regulatory friction that legal-tech for the poor consistently runs into.
A nonprofit building tenant-rights tooling — letter generators for repair requests, eviction-readiness tools, a bad-landlord database. Started in New York and now used by tenant organisers across multiple states. The model for what jurisdiction-specific civic legal-aid software actually looks like in production.
Document-assembly tooling for legal aid
A category rather than a single project. Document-assembly platforms (notably docassemble, the open-source legal-document assembler maintained by Jonathan Pyle) underpin a meaningful share of US legal-aid self-help interfaces. Listed as the substrate that legal-aid orgs build on, not as a single product.
The federally chartered backbone of US civil legal aid — funding the network of legal-aid organisations that take eviction, family-law, and benefits cases for clients who can't afford counsel. Pilot programs on AI in legal aid run through here; the institutional anchor the field organises around.
Voting access and election-information infrastructure. The research institutions and nonprofits that make voter registration, election administration, and election data legible to the public.
A nonpartisan law and policy institute at NYU School of Law. Long-running voting-rights research, election-security work, and litigation on redistricting, ballot access, and election administration. The most reliable institutional source for what the actual evidence shows on election integrity questions.
Center for Tech and Civic Life
A nonprofit serving local election officials with training, technology, and grants for election administration. The unglamorous but load-bearing work of supporting the county-level officials who actually run American elections day-to-day.
The largest nonpartisan voting-tools nonprofit in the United States — voter registration, polling-place lookup, ballot tracking. A working example of civic-tech infrastructure operating at the scale of real election cycles, with privacy and partisan-neutrality discipline.
Pew Charitable Trusts — election work
Pew has historically incubated open election-information infrastructure (the Voting Information Project) and conducted long-running survey work on election administration. Worth tracking as an institutional source for cross-state election data and the comparative research that election-integrity advocates rely on.
MIT Election Data and Science Lab
An academic lab producing and releasing election datasets — precinct-level returns, the Survey of the Performance of American Elections, election performance metrics. The empirical substrate that almost every serious quantitative claim about US elections eventually traces back to.
Public-data work on policing and the carceral system. Where official channels are unwilling or structurally unable to produce the data, these projects do the documentation work the public record otherwise wouldn't.
A comprehensive national database of police killings in the United States, compiled from public records, news reports, and direct submissions. Has filled the gap that the federal government's voluntary, fragmented police-killings reporting consistently leaves — and is the source of much of the per-capita data used in academic and journalistic work.
A searchable database of complaints, settlements, and use-of-force incidents involving Chicago Police Department officers. Built and stewarded by the Invisible Institute through years of FOIA litigation. The most-cited model for what officer-level public-records databases can look like when journalism and litigation work in parallel.
A Chicago-based public accountability journalism organisation. Long-running investigative work on policing, including Pulitzer-recognised collaborations on missing-persons cases and police misconduct. The institutional home of much of the city's police-records public-interest litigation.
A community-based research and abolitionist organising group focused on LAPD's surveillance infrastructure — gang databases, predictive policing, intelligence-led policing programs. Has produced some of the most detailed publicly available analysis of municipal-police data systems, grounded in directly affected communities.
Carceral Tech Resistance Network
A loosely connected network of researchers, organisers, and journalists working on carceral technology — electronic monitoring, prison communications, jail-management AI. Listed as a category because the field is decentralised by design; the work happens through shared research norms rather than a single institutional hub.
Capital aligned with public interest over corporate AI. The funder ecosystem most realistic as multi-year support for projects in this domain — observation, not endorsement, and worth checking against current strategy notes since funder priorities shift.
Long-running funder of journalism infrastructure, civic technology, and the institutions of an informed democracy. Active in civic-AI grantmaking and a major historical funder of the local-newsroom layer that algorithmic-accountability journalism depends on.
Funder of journalism, criminal-justice reform, and technology in the public interest. The Journalism & Media program and Safety & Justice Challenge have funded core elements of the algorithmic-accountability and decarceration ecosystems for over a decade.
A major social-justice funder. Programs in Future of Work, Technology and Society, and International Cooperation and Global Governance fund the digital-rights and accountability institutions that this domain rests on — including a meaningful share of the public-interest-tech grantmaking landscape.
The largest private funder of digital-rights work globally. Independent media, civil society technology, criminal-justice reform, and the cross-border human-rights infrastructure that underpins much of the international algorithmic-accountability community.
Funder of the Internet Health Reports, Mozilla Fellowships, and adjacent platform-accountability work. Smaller cheques than the larger foundations on this list, but with a track record of placing patient capital with researchers and projects before their work is institutionally legible elsewhere.
A research funder backing AI research with public-interest dimensions, AI safety, and adjacent work. Less embedded in the civil-society accountability ecosystem than the foundations above, but increasingly relevant as a source of research funding for groups working at the technical-policy boundary.
This page is a sample of the Justice & Civic domain. Inclusion is editorial — not a commercial endorsement, not a vetting of every claim a project makes, and not a recommendation that any single tool or organisation fits every situation. Foundations listed here are not partners; they're documented as the funder ecosystem most aligned with this work, public information, no relationship implied. If you build in this space and we missed you, submit your project below.
Beyond the catalog-wide exclusions, Justice & Civic has its own line. Tools that wear the language of safety, efficiency, or modernisation while operating as enforcement infrastructure against the people the catalog is built for are not the work. Naming the line is part of the work.
If your work fits the Justice & Civic domain — and is honest about which line it sits on — we want it in the catalog. Submissions are reviewed editorially before inclusion. We may reach out for clarifying details.
Pulled live from the AI for Planet catalog — reviewed editorially, with maturity and evidence flagged. Inclusion is not endorsement; see the curatorial note below. Submissions for the Justice & Civic domain open Q3 2026.