Urgent.News

What's breaking now, across thousands of outlets.

AI

Twitch faces a class action lawsuit over using streamers' content to train Amazon's AI

A Twitch streamer has filed a proposed class action against the platform and its parent company Amazon, claiming his broadcasts, and those of countless other creators, were fed into generative AI training with neither consent nor payment. Read more

Twitch faces a class action lawsuit over using streamers' content to train Amazon's AI

A Twitch streamer named Warren Pandiscia has filed a proposed class action lawsuit against the streaming platform and its parent company Amazon. Pandiscia claims that his and many other creators' broadcasts were used to train Amazon's AI without their consent or payment. According to the 37-page complaint filed in California, Twitch and Amazon copied large amounts of streamer footage without obtaining proper licenses, violating creators' implied agreements with the platform and state unfair competition laws.

The streaming giant announced its plans to use generative AI models earlier this month, leading to backlash from the Twitch community. Streamers demanded that the feature be optional, and the most popular request. Twitch's chief product officer, Mike Minton, explained that opting in would be too confusing for users, so the feature is enabled by default.

However, users can opt-out individually. The lawsuit comes after Twitch admitted that it uses creator content, including streams, past clips, chat logs, and photos, to train Amazon's AI models, with the opt-out not being retroactive. Pandiscia had no knowledge of his content being used for AI training and opted out as soon as the setting was available. Neither Twitch nor Amazon has commented on the lawsuit.

Written by urgent.news from Eurogamer's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at eurogamer.net →

More in AI

The Runaway Diff: A Token-Budget Postmortem for Coding Agents

The task looked trivial on paper: add a rate limiter to a small Python service and update three call sites. I handed it to a coding agent running on MonkeyCode, an open-source project with free model…

  • Engineer assigned a token allowance for coding agent on MonkeyCode platform
  • Diff output was two-thousand lines for change that should have been forty lines
  • Fix was a simple Python script imposing token budget and detecting edit oscillation

The Coverage Loop: Turning Free AI Tokens into Verified C++ Tests

Unit tests are boring. Coverage is not optional. I asked a free AI model to write my tests. It failed. Then I built a feedback loop.

  • AI model generates C++ tests for string utility
  • Coverage measured with gcov, loop repeats until 90% achieved
  • Feedback loop refines tests based on missing lines

More from Sunday 23 August →