Bluesky Suddenly Trending, Refuses to Train AI on Your Posts

The new policies of X, which permit exactly that, took effect last Friday.

Bluesky, a Privacy-Conscious Social Platform

Bluesky, the emerging social platform since the US election, announced on Friday that it does not use user posts to train its generative AI. This statement marks a clear distinction from the AI training policies of other platforms like X (Twitter) and Meta’s Threads.

Upholding User Data Integrity

According to The Verge, Bluesky stated: “We do not use your content to train generative AI, and we do not plan to.” Bluesky clarified that it employs AI to assist in content moderation, which helps sort posts and shield human moderators from harmful content. AI is also used in the algorithms powering its “Discover” feed.

A Non-Exploitative Data Policy

The Verge noted that Bluesky’s policy (robots.txt) does not prevent OpenAI, Google, or other GenAI companies from crawling its data. However, Bluesky justified this potential gap by highlighting the open and public nature of the platform. Bluesky spokesperson Emily Liu told The Verge: “We aim to do our part in ensuring that external organizations respect user consent, and we are actively discussing how to achieve this within our team.”

Significant Growth for Bluesky

Despite being an underdog compared to X and Threads, Bluesky has experienced significant growth following the US election. It surpassed the 15 million user mark last Wednesday, after gaining over a million new users in just one week. This growth coincides with an increase in account deactivations on X, as reported by SimilarWeb, which noted that “over 115,000 American web visitors deactivated their [X] accounts” on November 7, a record since Elon Musk took over the company. Concurrently, Bluesky’s web traffic and daily active users saw substantial increases the week before the election and again after election day.