
RIAA chief discusses AI training, deepfakes, and music copyright
The Decoder interview covers lawsuits, detection and labeling, and whether Spotify and YouTube would participate.
47 stories, newest first · one card per story

The Decoder interview covers lawsuits, detection and labeling, and whether Spotify and YouTube would participate.
Across three domains, six models and three frameworks, owner verification reached 99.2% and buyer Top-1 attribution 99.3%.

Publishers including The New York Times file lawsuit against Brave over scraped content shared with AI firms like Anthropic, Perplexity, Mistral, xAI and Cohere.

The reported limits cover modding and decompiling, and Anthropic has not explained the change.

In an amicus brief, Wikimedia says unlicensed LLM pre-training can be fair use, but query answers that replace the source need a license.
The post frames The New York Times copyright suit as seeking to discard longstanding fair-use doctrine for model training.
Authors report a scalable 32–128-bit payload, a 94.6% decode rate against CtrlRegen+, and 98.8% forgery detection.

Bartz paid about $3,000 per work; Cruz seeks up to $150,000 and an injunction on Claude training.
The authors say cheaper AI-assisted reverse engineering strains trade secret, copyright, anti-circumvention, misuse, and contract law.

Revised business terms still let Meta use site content to develop products, which lawyers say can include training.

The order also covers who approved downloading shadow libraries for Llama 1 and 2 and what happened to unused pirated works.

404 Media reports Reverse Engineer Anything passed 342,000 GitHub stars amid TV, IoT and software clone posts.

Developers describe a bind: sharing prototypes invites copies, while withholding them costs sales.
The terms take effect October 30, 2026, and the News/Media Alliance says business accounts cannot opt out.

Meta removed terms language that appeared to allow scraping business sites for AI training.

decomp.dev is shutting down after AI made game decompilation and cloning trivial, overwhelming human-led projects.

A reported Cartoon Network leak exposed over 200GB of production files, but its source remains unconfirmed.
AI decompilation is reconstructing commercial games, aiding preservation while raising fears of software reuse.

Claude Opus 5.5 is decompiling classic games into working browser ports, complicating IP enforcement.

USA Today Co. and its local newspapers sue OpenAI for over $250 million, alleging unauthorized copying of hundreds of thousands of articles for AI training.
South Korean President Lee called for AI-powered cyber defenses after data leaks hit Coupang and local banks.

Suno's new Speech feature generates spoken audio and matching background music as one track.

Google's pilot paying publishers for AI Overview contributions is struggling amid tiny, opaque payouts and lawsuits.

Opinion essay says OpenAI agents attacked websites and scraped paywalled news, arguing executives should face consequences.
Today 1.5K readers · 2.4K views · 754 reads · all time 189K views