Bluesky appeals to weary social media users with its AI policy. Credit: Hakan Nural / Anadolu / Getty Images
Bluesky, the X competitor that's attracted over three million followers since the U.S. presidential election, doesn't train generative AI models on user data.
In a post on Friday, Bluesky said, "we do not use any of your content to train generative AI, and have no intention of doing so." In a follow up post, it explained that it uses AI to assist in content moderation and in the Discover algorithmic feed, before adding "None of these are Gen AI systems trained on user content."
The announcement coincided with major changes on X and an influx of new users to Bluesky. Recently X changed how its block function worked and modified its privacy policy, allowing its LLM Grok to train on user data. This, plus Elon Musk's vocal support of President Trump appears to have driven an exodus from X -- an many are turning to Bluesky.
Bluesky currently has over 17 million users, jumping up from 9 million users as of September. Users have grown to mistrust companies like X, Meta, and Google for using their data to train generative AI models without ways of opting out. Since Bluesky has become perceived as the less toxic version of X, the company's explicit statement about its user data policies is welcome news for users sick of their data being exploited.
That said, Bluesky doesn't currently have any generative AI features, so it's easy for the app to say it isn't training anything. As we know from X's frequent policy updates, all that could change. Bluesky's Terms of Service, which was linked in the post doesn't have any explicit mention of AI model training, so the downside is that a new clause could easily appear. In other words, never say never, but Bluesky users are safe for now.