A streamer has initiated a class action lawsuit against Twitch and its parent company Amazon, asserting that their content was used to train artificial intelligence models without proper permission or compensation. The lawsuit, filed by Connecticut-based streamer Warren Pandiscia, centers on Twitch's recent program that utilizes streamer content for AI training.
According to the 37-page complaint, Pandiscia alleges that Amazon had a significant incentive to gather training data on an unprecedented scale for its commercialized AI products. Instead of securing lawful licenses or seeking explicit permission, the suit claims that Twitch streams and videos were accessed and utilized as a substantial dataset to fuel Amazon’s AI initiatives. "Rather than negotiate for lawful licenses or seek permission, defendants accessed the Twitch streams and videos to utilize them as a massive dataset necessary to fuel Amazon’s AI products," the complaint states.
Twitch's chief product officer, Mike Minton, reportedly acknowledged the unpopularity of the AI policy, admitting that "nobody would opt in" if the default setting were not automatic. However, the lawsuit raises concerns about potential loopholes, suggesting that even with the option to disable data usage on one's own channel, content shared by others could still be harvested.
Further alarm is sounded by the possibility that Amazon may have retroactively utilized content from before the opt-out option was available. Citing previous comments from Minton, Pandiscia's suit claims that Twitch has the capability to use videos dating back to early 2024, regardless of whether a streamer has opted out. "Because Amazon AI products are commercialized, Amazon had an overwhelming incentive to acquire training data on an unprecedented scale," the suit alleges.
As concerns grow around the quality and ethical sourcing of AI-generated content, the lawsuit against Twitch and Amazon could set significant legal precedents. Pandiscia is seeking not only damages but also injunctive relief, which would compel the removal of his materials from AI training datasets. A successful outcome could have implications extending beyond Amazon's operations, potentially challenging broad practices of AI data harvesting.