Social Network Bruinsky Recently Posted suggestions on Github A new option is outlined, which allows users to point out whether they want their posts and data to be used to generate things like AI training and public archives.
CEO Jay Graber The proposal was discussed earlier this week Although on the southern stage in the southwest, on Friday night, it attracted new attention Posted information about it on the blues. Some users are full of alerts about the company’s plans, which they believe are a reversal of what the Blues previously insisted on. User data will not be sold to advertisers and Will not train AI on user posts.
“Oh, hell no!” user sketch Write. “The advantage of this platform is that there is no sharing of information. Especially the AI generation. You are not in trouble now.”
Grab your hands answer The generated AI companies “have scratched public data from the web”, including Blues, because “everything on the Blues is public like a website.” So, she said the Bluesky tried to create a “new standard” to manage this scratch, similar to robots.txt Submit the file used by the website to convey its permissions to the network crawl.
Debate on AI training and copyright Drag Robots.txt into the spotlightamong other matters, emphasizes the fact that it is legally unenforceable. Bluesky framed its proposed standards as a standard with similar “mechanisms and expectations”, providing a “machine-readable format that is expected to be followed by the actors and does have moral weights, but is legally not enforceable.”
According to the proposal, users of Bluesky apps or other applications that use the basics Atrotocolcan go into their settings, allowing or not allowing their blues data to be used in four categories: Generate AI, protocol bridge (i.e. connecting different social ecosystems), batch data sets, and network archives (such as Wayback Machine for Internet Archives).
The proposal says that if users say they do not want the data used to train the AI to generate, “the company and research teams either respect that intention when they see its intention when they see it or do batch transfers on the scraping site or using the protocol itself.”
Molly White Describe this As “a good advice” and say “it’s weird to see people burn Bruinsky for this” because it’s not “welcome to AI scratches” but rather “trying to add consent signals to allow users to communicate preferences for scratches that have already occurred.”
“I think this weakness and [Creative Commons’] A similar proposal for “preference signals” is that they rely on scrapers to respect certain desires to be good actors,” White continued. “We have seen some of these companies blow over robots. txt or pirate material to scratch. ”
