Bluesky will allow you to configure consent for the use of user data to train AI

Bluesky has indicated that he is working on a configuration option with which to allow users to specify whether they want to give their consent for third-party developers to use your publications and information for train Artificial Intelligence (AI).

The ‘microblogging’ social network recently explained its position on generative AI, ensuring that does not train its models with user datanor does it have “intention to do so” in the future.

Specifically, he detailed that, in order to clear the doubts of users who have “concerns about the training of other platforms” with their information, the platform does not use “none” of its subscribers’ content to train generative AI. However, uses this technology to “assist in content moderation” and for the publications of the ‘Discover’ algorithmic ‘feed’.

Now, Bluesky has shared an update on the use of generative AI on the platform, where he has indicated that he is investigating an option configuration with which allow users to specify whether they consent so that external developers can use your content on AI training data sets.

As explained in a thread of publications on the platform, as it is a public and open social network, works “very similar to Internet websites”. Thus, it has been pointed out that these websites They can specify whether they agree to let third-party companies track their data with, for example, a robot.txt, in order to be able to use them later for other purposes, such as AI training.

In this sense, Bluesky is investigating “a similar practice” to that of websites, related to allowing external companies track the data and content published by users on the platform.

As you specified, with this configuration option the Users will be able to specify whether they give permission or not for third-party developers to use your content. However, Bluesky has also stressed that “you will not be able to enforce this consent outside of your systems.”

That is, although the user indicates that they do not allow the use of their content to train AI, the company has clarified that It will ultimately be up to third-party developers to “respect these settings.”

At the moment, Bluesky is “maintaining ongoing conversations with engineers and lawyers” to finish developing this information use configuration option. So more related information will be shared soon.

This novelty in the use of data for AI training coincides with the publication of a Hugging Face employee, in which Share information from a million Bluesky postsextracted from their API in the IA repository.

As Hugging Face explained, this publication provoked “a lot of criticism from the community” about their creation and uploading, as well as about the use of user data. Therefore, the company ended up removing Bluesky data from the repository and explaining that, although its intention was to “support the development of tools for the platform”, it is a approach that “violates the principles of transparency and consent in data collection.”