The Knight First Amendment Institute at Columbia University, also referred to as Knight Columbia, is a nonprofit institute working on the First Amendment, technology, and democracy. It was founded in 2017 with more than $30 million from the Knight Foundation and Columbia University, and is located at Columbia University in New York. The institute describes its methods as strategic litigation, research, policy advocacy, and public education (Source: knightcolumbia.org).
Essay series and research
Knight Columbia publishes the Knight Columbia essay series, a peer-edited venue for long-form essays on technology, democracy, and the public sphere aimed at an interdisciplinary academic and policy audience. Its AI output is organized under an "Artificial Intelligence and Democratic Freedoms" project studying how advanced AI systems may harm or strengthen democratic freedoms, which combines events, research and litigation (Source: knightcolumbia.org). The project grew out of a symposium held at Columbia University and online, announced January 21, 2025 for April 10–11, 2025 sessions, following a call for abstracts issued August 15, 2024 (Source: knightcolumbia.org). Seth Lazar joined the institute as Senior AI Advisor on August 2, 2024 (Source: knightcolumbia.org).
In 2025-26 the series carried several essays on AI governance. Its May 2026 symposium on AI in democratic society produced or republished five foundational essays:
- Anticipatory AI Ethics (Lazar, Knight Columbia, May 1 2026) (Lazar — Anticipatory AI Ethics; May 1)
- AI as Social Technology (Farrell + Shalizi, Knight Columbia, May 11 2026) (Farrell and Shalizi — AI as Social Technology; May 11)
- A Conceptual Model to Guide AI Risk Governance Strategies (Mulligan + Marda + Wang, Knight Columbia, March 16 2026) (Mulligan, Marda, and Wang — A Conceptual Model to Guide AI Risk Governance Strategies; March 16)
- Building AI for the Democratic Matrix: A Technical Research Agenda for Normative Competence and Normative Institutions (Hadfield + Trivedi + Hadfield-Menell, Knight Columbia, March 3 2026) (Hadfield, Trivedi, and Hadfield-Menell — Building AI for the Democratic Matrix; March 3)
- Levels of Autonomy for AI Agents (Feng + McDonald + Zhang, Knight Columbia, 2026) (Feng, McDonald, and Zhang — Levels of Autonomy for AI Agents; July 2025, republished May 2026)
The AI as Normal Technology essay (Narayanan and Kapoor, 2025), summarized at AI as Normal Technology, also appeared through Knight Columbia. Its Deep Dive series carried the same authors' follow-on reply, "Do AI Risks Require Extraordinary Government Intervention?", on May 21, 2026, which defines extraordinary intervention by three factors and argues for societal resilience over AI nonproliferation (Do AI Risks Require Extraordinary Government Intervention?).
These five 2025-26 essays share several characteristics: each rejects the AGI and Singularity framing as the dominant analytical primitive; each emphasizes the deployment environment as central to policy analysis; and each offers a named, citable framework rather than diffuse critique.
Further AI pieces in the series and the associated Deep Dive channel include Sayash Kapoor and Arvind Narayanan's Do AI Risks Require Extraordinary Government Intervention? (May 21, 2026), a reply to Derek Thompson arguing that resilience investment is preferable to precautionary restrictions on model release; Michael C. Dorf's Supporting and Implementing Truth as a Free Speech Value (February 19, 2026); The AI Power Disparity Index by Rachel M. Kim, Blaine Kuehnert, Lazar, Ranjit Singh and Hoda Heidari (September 8, 2025), proposing a composite indicator for the distribution of power in the AI ecosystem; Lazar and Mariano-Florentino Cuéllar's AI Agents and Democratic Resilience (September 4, 2025); Henry Farrell and Hahrie Han's AI and Democratic Publics (August 1, 2025); Can AI Mediation Improve Democratic Deliberation? by MH Tessler and eleven co-authors (August 1, 2025); Towards Interactive Evaluations for Interaction Harms in Human-AI Systems by Lujain Ibrahim, Saffron Huang, Umang Bhatt, Lama Ahmad and Markus Anderljung (June 23, 2025); Daniel Susskind's What Will Remain for People to Do? (April 7, 2025); Nadine Farid Johnson's Responsible AI Regulation: Supporting Independent Researchers (March 14, 2025), on coordinated flaw reporting for general-purpose AI; and A Safe Harbor for AI Evaluation and Red Teaming by Shayne Longpre, Kapoor, Kevin Klyman, Ashwin Ramaswami, Rishi Bommasani, Narayanan, Percy Liang and Peter Henderson (March 5, 2024), arguing for legal and technical safe harbors for AI safety and trustworthiness research (Source: knightcolumbia.org). The institute also convened a February 20, 2024 symposium, Generative AI, Free Speech, and Public Discourse.
Litigation and advocacy
The institute's AI work extends beyond publication into litigation and amicus practice, generally on the side of disclosure and research access rather than developer speech claims.
On July 22, 2026 the institute urged the Ninth Circuit to reject xAI's bid to block the California law requiring disclosures about generative AI training data, under the heading that the public needs greater transparency into how generative AI systems are built (Source: knightcolumbia.org). The California measure at issue is AB 2013. The institute has also produced an opinion analysis of Amazon v. Perplexity AI, a case addressing the application of computer crime laws to AI tools that collect information on behalf of users (Source: knightcolumbia.org).
Earlier technology-surveillance work includes a March 12, 2025 action with the SMU Law Clinic seeking immediate release of records related to a Texas school district's use of surveillance technology, on the argument that school surveillance systems undermine students' privacy and expressive rights (Source: knightcolumbia.org). In October 2023 the institute joined public-interest organizations in calling on Congress to consider the impact of AI on civil society.
Relationships
- affiliated-with: Columbia University
- funded-by: Knight Foundation
- publishes: Anticipatory AI Ethics (Lazar, Knight Columbia, May 1 2026), AI as Social Technology (Farrell + Shalizi, Knight Columbia, May 11 2026), A Conceptual Model to Guide AI Risk Governance Strategies (Mulligan + Marda + Wang, Knight Columbia, March 16 2026), Building AI for the Democratic Matrix: A Technical Research Agenda for Normative Competence and Normative Institutions (Hadfield + Trivedi + Hadfield-Menell, Knight Columbia, March 3 2026), Levels of Autonomy for AI Agents (Feng + McDonald + Zhang, Knight Columbia, 2026), AI as Normal Technology, Do AI Risks Require Extraordinary Government Intervention?
- litigates: xAI LLC v. Weiser (challenging the Colorado AI Act) — related California training-data disclosure litigation in which the institute filed against xAI's Ninth Circuit challenge; Amazon v. Perplexity AI — subject of an institute opinion analysis.
- related: Lawfare (parallel venue for legal-and-policy essays), AI as Social Technology, Anticipatory AI Ethics (Lazar framework, including the Technological Horizon), Sociotechnical AI Risk Governance, Normative Competence, Levels of Autonomy for AI Agents (Feng-McDonald-Zhang framework), AI as Normal Technology
- related: California AB 2013 — Generative AI Training Data Transparency — the training-data disclosure law it defended at the Ninth Circuit; AI Transparency — the disclosure position running through that filing and the safe-harbor essay.
Sources
- Multiple foundational source pages (see "publishes" above)
- Knight First Amendment Institute, artificial-intelligence tag index: knightcolumbia.org
- Knight First Amendment Institute, "Artificial Intelligence and Democratic Freedoms": knightcolumbia.org