Bluesky won’t use your posts for AI training, but can it stop anyone else?

Today, a Hugging Face employee published data from 1 million Bluesky posts scraped from its API to the AI repository. He’s removed it and apologized, but 404 Medianotes the set was “trending” all day.

Bluesky says it’s looking into ways to “specify consent (or not) for AI training.” but acknowledges that “It will be up to outside developers to respect these settings.”