Social network Bluesky announced on Friday that it has “no intention” of using user content to train generative AI tools. This statement came on the same day X’s updated terms of service, allowing user text and information to be used for training its AI tools, went into effect.
Bluesky addressed creators’ concerns, stating in a post: “Many artists and creators have made Bluesky their home, and we understand their concerns about other platforms training on their data. We do not use your content to train generative AI and have no plans to do so.”
However, it acknowledged that other companies could potentially scrape Bluesky posts for training, as its robots.txt file does not block crawlers from companies like Google or OpenAI. Bluesky spokesperson Emily Liu noted, “Bluesky is an open and public network, similar to websites on the Internet. While robots.txt files don’t always stop companies from crawling, we aim to ensure that outside organizations respect user consent and are discussing ways to achieve this.”
A number of artists and creators have made their home on Bluesky, and we hear their concerns with other platforms training on their data. We do not use any of your content to train generative AI, and have no intention of doing so.
Bluesky clarified that while it uses AI for tasks like content moderation and improving its algorithmic Discover feed, these systems are not generative AI trained on user content. The platform provided a link to its terms of service, community guidelines, and other policies for transparency.
Over the past week, Bluesky has grown by more than three million users, reaching a total of 17 million. This growth has brought challenges such as increased spam, scams, and trolling, prompting the company to expand its moderation team.
Bluesky’s rise coincides with increased competition among microblogging platforms. Meta’s Threads, a major competitor, recently reported over 15 million signups this month and announced testing of custom feeds—a feature Bluesky is known for. Unlike Bluesky, Meta has admitted to training its AI models on nearly all publicly shared content since 2007.
Correction, November 15th: It was previously reported that Bluesky made its statement before X’s new terms of service took effect, but those terms are already active as of today.
