Bluesky users debate plans around user data and AI training

Bluesky recently social networks Published a proposal on Gut Hub Outline the new powers can indicate that they want them to scare their posts and data for things like Generative AI training and public archives.

CEO J. Graber discussed the proposal earlier this week, while on the stage southwest of the southwest, but then, on Friday night, he turned the latest attention. Posted on Bluesky about this. Some consumers reacted with the alarm to the company’s plans, which they saw as Bluesky’s former insistence that they would not sell user data to advertisers and would not train AI on user posts.

“Oh, no hell!” The user’s sketch Is written. “The beauty of this platform was not sharing information. Especially General AI. Now you don’t do cave. “

Grabber Replied These generative AI companies are “already eliminating public data from the web”, including Bluesky, because “everything on Bluesky is public is like a website.” So he said Bluesky is trying to create a “new standard” to rule over such a scratch, as is like the Lord Robotes DotTS Files that websites use to convey their permission to web crawls.

Discussions about AI training and copyright have dragged robots. Bluesky makes his proposed standard a similar “mechanism and expectations”, which “provides a format for reading machine, which is expected by good actors, and he lifts moral weight, but it is not legally viable.”

Under this suggestion, the Blueski app usage, or other apps that use the basic atprotocolThey can go into their settings and allow or allow the use of their Bluesky data in four types: Generative AI, Protocol Burgenging (ie connecting different social environmental systems), bulk datases, and web archives (such as Internet archive back machine).

If a user indicates that they do not want to use their data for generative AI training, the proposal states that “AI training set -making companies and research teams are expected to be used when they scratch the website, or use the protocol themselves.”

Molly White, who writes the reference is needed newsletter and the web 3 is just going to be a great blog, Described it As a “a good suggestion”, and said, “It’s strange for people to see Bluesky for it,” since it is not so “welcome to AI scraping” but “trying to add a signal to the consent so that users are allowed to communicate with priorities for scraping.”

“I think his weakness and [Creative Commons’] White added, a similar suggestion for ‘priority indicators’ is that he relys on scratching to respect these gestures by desire to become a good actor. “We have already seen that some of these companies have been blown up for past robots dot TX or seafood content.”

Leave a Comment