hearing of ai moderation going brrrrr towards users, suspending them but then eventually getting unsuspended with a final warning also falsely accusing them of interacting with certain seriously bad content.
Comments
Log in with your Bluesky account to leave a comment
like it or not, ai moderation *has* to be used in order to moderate posts at large scale. users should however be able to dispute automated suspensions or warnings & actually have a person review their case. this, also in hopes that the dataset could also be improved to further avoid false positives
I’m assuming that if person A gets hit from interacting with someone’s post and person B happened to interact with any of person’s A posts then they get hit too. if that’s genuinely the case, it makes sense *but* also the way they’re doing the chained suspensions is terrible and risky.
Comments