Bluesky says it won't train generative AI on your posts on the platform

Bluesky has been gaining popularity recently and now, it's saying it won't train generative AI on your posts on the platform. That differs from the approach its competitor X (formerly Twitter) is taking, as well as Meta's Threads.


Bluesky has said they are not using any of your content to train AI and have no intention of doing so. The decentralized social platform has subsequently clarified that it uses AI for content moderation, and also in the algorithm powering its Discover feed.

However, the company underlined that neither the content moderation AI nor the one powering its Discover feed is trained on user content.

The Verge points out that Bluesky's policy that dictates what outside party can scrape from a website doesn't prevent companies like OpenAI, Google, or other companies involved with generative AI from crawling its data.

Bluesky spokesperson Emily Liu said that the company is doing their part to ensure that organizations respect user consent. Liu also said Bluesky is actively discussing with the team how to achieve this while maintaining Bluesky as an open and public social network just like any website on the Internet.

Bluesky is still not as big as X or Threads, but it has been gaining popularity recently and has surpassed the 15 million user threshold back on Wednesday last week. It's added more than a million users in the past week and is working on becoming a good alternative to X and Threads.

Bluesky seems to be placing itself in a good position to rival X and Threads, and the fact that it promises it won't train generative AI on your posts is a plus.

X's option should allow you to opt out of training AI, but I still feel like this should rather be an "opt-in" feature while being off by default, instead of the other way around. So I do like Bluesky's commitment here.