Bluesky users discuss the plans about user data and artificial intelligence training

Social Network Bluesky recently Publish a proposal on Gitap Determining new options can give users to indicate whether they want to embody their posts and data for things such as artificial intelligence training and general archiving.

CEO Jay Ghubar Discuss the proposal earlier this week, While on the stage in the south of the southwest, it attracted new attention on Friday night, after that Posted on Bluesky. The reaction of some users to warning the company’s plans, which they saw as a reflection of the previous Bluezki insistence that it was You will not sell user data to advertisers and You will not train artificial intelligence on user publications.

“Oh, hell no!” User drawing books. The beauty of this platform was not to share information. Especially General AI. Do not stand now. “

Grapher to reply Trucket intelligence companies are “already a general data coding from all over the web”, including from Bluesky, because “everything on Bluesky is the year like a general web site.” She said that Bluezi is trying to create a “new standard” to judge this bulldozing, similar Robots.txt The file that web sites use to connect their bills to the web crawling.

Discussions on artificial intelligence training and copyright Robots.txt withdrawal in the spotlightAmong other things, highlighting the fact that it is not legally implemented. Bluesky develops its proposed standard as a similar “mechanism and expectations”, and provides “readable coordination, which is expected to be committed by good actors, and bears moral weight, but it is not legally implemented.”

Under the proposal, Bluesky users or other applications use the foundation ATPRTOCOLIt can be included in their settings and allowing or not allowing the use of their Bluesky data in four categories: the Touly AI, the protocol bridge (i.e. the connection of different social environmental systems), wholesale data groups, and web archiving (such as Internet Archive Web).

If the user indicates that they do not want to use his data to train artificial intelligence, the proposal says: “It is expected that it will respect companies and search teams that build this intention of Amnesty International when they see them, either when stripping websites, or wholesale transfers using the protocol itself.”

Molly White, who writes the required newsletter and web3 will make a great blog, Describe As a “good suggestion”, and he said “it is strange to see people who move on Bluesky for him”, because he does not welcome much in the scraping of artificial intelligence, “but” trying to add an approval signal to allow users to connect preferences to the already being ran. “

“I think weakening with this and (creative rumors”) is a similar proposal for “preference signals” is that they depend on the disciplines to respect these signals from some desire to be good actors, “White continued.