The mention of API does make me think: can we create files similar to robots.txt to tell AI systems that whole domains and their hosted content is off limits? That would abolish needing to make yet another account for self-hosted sites? #AI @spawning_ do you know?
Spawning
 @spawning_
Mar 7
Replying to @spawning_
We partnered with @ArtStationHQ to ensure opt out requests that have been made on their site are honored in our API, and will not be present in future AI model training with participating organizations.
1
1
we are collaborating on something very similar to be available very soon. We can do whole domain opt-out by default.
1
2
Maybe @w3c & @w3cdevs want to get in on a web standard for AI crawling & collection as well if they aren't already? Formalizing a standard and developing headers etc to let webmasters & front end developers define boundaries for AI/ML systems seems like a logical step.
2
1
Replying to @Aerisnoir @spawning_ @w3c
There is no active standardization work in this space, nor have we identified pre-standardization or incubation work. If indeed a proposal emerges, bringing it to @WICG_ or in a new #W3C Community Group would be a good first step towards standardization w3.org/community/

Mar 8, 2023 · 1:00 PM UTC

1
2
Replying to @w3cdevs @Aerisnoir @w3c @wicg_
we are contributing to work happening and I'll share this info amongst the group
1
1
I'm reading things in a C that are somewhat reassuring. I have a hunch you know what I'm reading 😁