Monitor this situation.

Unsubscribe anytime.

[SITUATION] · [QUIET] · [TECHNOLOGY]

2 clusters · 6 sources · 3 days · First seen 27 days ago · Last updated 25 days ago

Social media platforms using user data for AI training

Overview

Social media platforms are increasingly utilizing user-generated content to train generative artificial intelligence models. This trend is characterized by a shift toward implicit consent models, where platform usage is often treated as permission for data harvesting.

Twitch, owned by Amazon, implemented a policy that enables the use of live broadcasts, video recordings, and chat messages for AI training by default. A Twitch product manager noted that an opt-in system would likely result in “insufficient consent levels to meet AI development needs.” While Twitch allows users to manually opt out via privacy settings, many other social media services either make the opt-out process difficult or do not provide such a mechanism at all.

While some platforms like Bluesky do not officially train models on user data, their public protocols remain susceptible to third-party scraping. Meanwhile, major entities such as Google and Facebook continue to leverage user data for various AI-driven functions.

Entities

Twitch · Amazon · Google · Facebook · Bluesky

Timeline

25 days ago

[TECHNOLOGY] 2 sources
Social media platforms use user content for AI training
Many social media platforms use user content to train generative AI models, often without clear opt-in consent or easy opt-out settings, though some services like Twitch have introduced user controls.
27 days ago

[TECHNOLOGY] 4 sources
Twitch enables default use of user content for Amazon AI training
Twitch has implemented a default opt-out policy allowing Amazon to use live streams, recordings, and chat data to train generative AI models.

Sources

bigbaddice.pl · gadgetgear.nl · ihal.it · lifehacker.com · lifehacker.com.au · yesky.com