External ideas, open to examination

Source discussions.

Explore a publication as its own discussion: its perspectives, claims, questions, and connections to other ideas.

The reading collection

6 publications

01Imported publication · June 2025

AI as Governance

Henry Farrell

The abstract proposes two ways for political scientists to study AI: as a tool used within governance, or as a way of organizing and coordinating people in its own right. It suggests these approaches could raise new questions about markets, bureaucracies and democracies, but it reports no demonstrated political effects; the full article was not available.

Source coverage

Abstract only. The full article could not be retrieved; no claims about its full argument are inferred.

Open the discussion 6 connected ideas

02Imported publication · 18 October 2024

AI can help humans find common ground in democratic deliberation

Michael Henry Tessler, Michiel A. Bakker, and colleagues

The article reports UK experiments with 5,734 participants who discussed issues with help from an AI that drafted and revised group statements using their opinions and critiques. Participants often preferred the AI’s statements to statements written by human mediators, and groups became more aligned after AI-assisted discussion than after simply viewing one another’s opinions. The results do not show that the approach will work in other countries or larger groups; the tested assembly also left out fact-finding and expert testimony, and the AI could not check facts or moderate discussion.

Source coverage

Full readable text of the 10-page Science article copy, including references. Supplementary materials are not included.

Open the discussion 114 connected ideas

03Imported publication · December 2023

Fostering More Inclusive Democracy with AI

Hélène Landemore

The article argues that people and governments can shape AI, and proposes giving citizens a voice alongside experts and regulators. It suggests AI might help more people take part in public discussion, but reports that, as of December 2023, there was no evidence that AI facilitators did better than human facilitators. The author also warns about risks such as bias, privacy loss and excluding people who cannot read well or distrust technology, and calls for political, legal and social action as well as safe tools.

Source coverage

Selected article excerpts: opening argument, December 2023 evidence caveat, risks and conclusion. Full-document coverage is not claimed.

Open the discussion 16 connected ideas

04Imported publication · 2024 preprint; retrieved version recorded in source snapshot

Large Language Models (LLMs) as Agents for Augmented Democracy

Jairo F. Gudiño, Umberto Grandi, and César Hidalgo

This research paper tests six adapted language models using policy choices from 267 volunteers in an online project during Brazil’s 2022 election. The authors report that the models predicted choices better than a simple rule based on political orientation, and that adding model predictions improved estimates of the volunteers’ combined preferences in some tests; accuracy varied across groups, and reversing the order of options sometimes changed predictions. The study does not show how people across Brazil would vote or whether AI helpers would work safely in real decisions, which the authors say these models are not ready to make.

Source coverage

Full readable text of the retrieved 37-page preprint, including references and appendices. Figures are not independently interpreted.

Open the discussion 124 connected ideas

05Imported publication · Living book; chapter snapshot imported 17 October 2026

Plurality — Augmented Deliberation

Audrey Tang, E. Glen Weyl, and the Plurality Community

The chapter describes ways people can use conversation, writing, and digital tools to share views and make decisions. It presents Community Notes, Polis, and other projects as examples, while distinguishing reported uses and study findings from proposals that have not been tested. It also discusses limits such as slow discussion, unequal participation, and the difficulty of checking how AI systems reach their answers.

Source coverage

Complete Augmented Deliberation chapter from the living book. Other chapters are not included.

Open the discussion 83 connected ideas

06Imported publication · 2016; author manuscript

The Threat of Algocracy: Reality, Resistance and Accommodation

John Danaher

Danaher argues that public bodies may rely more on computer programs to organize information and shape decisions. He warns that people could lose the chance to understand or take part in decisions, even when a human remains involved. The article weighs possible benefits against this risk and considers responses, but it does not show that any proposed response has solved the problem.

Source coverage

Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version.

Open the discussion 144 connected ideas