Living dossier
generative AI
A living Archypedia dossier connecting source-backed reporting, context and new developments about generative AI.
Evidence timeline
Coverage as it developed
Newest first. Every entry retains its own source ledger and publication history.
-
Sony Music files second lawsuit against Udio over 30,000 recordings
The legal action follows discovery in a separate case, with Sony claiming Udio bypassed protective measures to train models on proprietary sound recordings.
Technology · 603 words -
Google AI search poses unacceptable risks to children, report finds
A report by the Youth AI Safety Institute documents failures in Google's AI search tools regarding crisis detection, content accuracy, and academic integrity. The institute recommends that Google disable these features by default for all minor accounts.
Technology · 628 words -
Suno scraped millions of songs from YouTube and other platforms via hack
A security breach at Suno has confirmed the systematic scraping of millions of audio files and lyrics for training its AI models. The leaked internal logs provide new evidence for record labels currently suing the company over alleged copyright infringement.
Technology · 534 words -
Meta users can now opt out of allowing AI to reuse public Instagram photos
Meta has introduced a feature that uses public Instagram photos as social context for AI-generated images. Users can adjust their privacy settings to restrict how their content is utilized by the tool.
Business · 657 words -
Viber integrates ChatGPT to add AI tools directly into messaging chats
Rakuten Viber has officially integrated ChatGPT into its messaging platform, enabling users to access AI tools for translation, text editing, and image remixing.
Technology · 637 words -
Google updates privacy settings to use user media for AI training
Google has updated its privacy policies to automatically collect images, files, and audio for AI training. Users can manage or disable this via their account.
Technology · 573 words -
Google launches Nano Banana 2 image model with high-speed generation
Google’s new Nano Banana 2 image model bridges the gap between high-speed iteration and studio-grade quality across its ecosystem. The release includes a cost-effective Lite version and enterprise-ready features like SynthID watermarking.
Technology · 652 words -
Tidal stops royalty payments for fully AI-generated music
Tidal has announced a new strategy to demonetize music fully generated by synthetic tools while introducing AI-labeling badges to improve transparency.
Technology · 606 words