Bluesky users debate plans around user data and AI training

Social Network Blusky recently Reveal an offer on Githab By creating the outline of new options, it may indicate that users should scrap their posts and data for things like generator AI training and public archive.

CEO J Grareer Discussed the proposal at the beginning of this week, While in the south-south-west level, but on Friday night it attracted new attention, after that Posted about it in Blusky about itThe Some users responded to alarms in the company’s plan, which they saw as the opposite of Bloosky’s previous Jedd Will not sell user data to advertisers And Do not train AI to the user’s postThe

“Oh, no hell!” User sketches WroteThe “The beauty of this platform was to share information. Especially Jenner AI. You don’t cave now. “

Gruber Replied Those generators AI companies with blusky “already scraping public data across the web”, since “everything on Bluiceki is public as public.” So he said that Blusky is trying to create a “new value” to manage that scraping, similar Robots.Text Website File that uses to communicate their permissions on web crawlers.

There is controversy about AI training and copyright Drag Robots.Text on the spotlightAmong the other issues it highlights the fact that it is not legally enforceable. Bluesky frames its proposed values ​​that will have the same “process and expectation”, “a machine-spaceable format, which is expected to adhere to good actors, and carry moral weight, but it is not legally applicable.”

Under the proposal, users or other applications of the blusky app that use the underlying AtprotocolThey can go to their settings and allow or disable their blusky data in four categories: Generator AI, Protocol Bridget (ie connects various social ecosystems), bulk datasate and web archive (such as internet archive’s weback machine).

If a user indicates that they do not want to use their data generator for AI training, the proposal says, “Companies and research teams are expected to respect this intention when creating AI training sets, while websites are scrapping websites, or when using the protocol itself.”

Molly White, who wrote the quotation required newsletter and web 3 simply going to great blogs, Described it As “a good offer” and said that it was “Blusky shining” for “it”, “since it is so” not to “welcome to AI scraping”, but “trying to add a consent signal to allow users’ preferences for scraping already happening.”

“I think this and the weakness with it [Creative Commons’] Similar proposals of ‘preferred signal’ are that they rely on scrapers to respect these signals from some of the best actors, “White said more.” We already saw some of these companies blowing up some of the past robots t