AI Comedy Copies Expose the Gap Between Labels and Authorship
Comedians' new complaints reveal why synthetic-media disclosure, recommendation rules and permission to reuse a performance require separate assessments.
Share
Copyright & Misinformation
35 posts
OpenAI’s EU Text Watermark Plan Highlights Limits of AI Authorship Detection
OpenAI plans EU text watermarking for ChatGPT and Codex, while acknowledging that edits weaken detection and neither a positive nor negative result settles human authorship.
Share
Anthropic Settlement Claims Put Publishing Rights Records Under Pressure
Authors are disputing publisher and agency claims on Anthropic settlement payments, exposing the difficulty of applying old rights records at scale.
Share
Pangram’s Rise Makes AI Detection a High-Stakes Editorial Tool
Publishers and platforms are adopting Pangram's probability scores, but false positives, bias and easy evasion complicate their use as evidence.
Share
US Government Backs OpenAI’s Fair Use Position in Copyright Case
A 20-page brief argues that restricting unlicensed AI training could impede innovation, but the filing is advocacy rather than a court ruling.
Share
Instagram Will Reduce the Reach of Undisclosed AI Profiles
Instagram is renaming its synthetic-profile label and says accounts presenting AI-generated people without disclosure may lose distribution.
Share
Music Publishers Accuse Anthropic of Widespread Copyright Theft
Sony Music Publishing, Warner Chappell and other publishers allege Anthropic used illegally acquired books containing lyrics and sheet music to train Claude.
Share
Cara Scraper Joins Artists to Track Stolen Work
After scraping 12 million works from Cara, a student apologized and began helping the artist platform build an open-source dataset monitoring tool.
Share
AI Training on Copyrighted Books Remains a Divided Legal Question
US rulings are separating lawful AI training from unlawful acquisition while fair use, market competition and human authorship remain unsettled.
Share
Claude’s Invisible Watermarks Face Workarounds Before Detection Launches
Developers are circulating methods intended to remove Anthropic’s invisible text watermarks, exposing the limits of machine-readable labels before its public detector is available.
Share