@scanlime @analogist @ariadne @timnitGebru

any tech solution /must/ have human oversight, otherwise its just going to cause more problems and maybe create further attack surface (it could end up rather like replacing the door security at a music venue with CCTV and robot "bouncers", which isn't likely to work well)

@vfrmedia @scanlime @analogist @ariadne @timnitGebru

The most popular federated block lists on twitter were intentionally used to target trans people, which is why we don't have them here. I think, because DARVO, tools like that, if deployed here, would quickly be subverted by racists. I would not use an instance that relied on such a list.

Indeed, i'd quit any instance that didn't moderate properly and take steps to protects its users. I strongly feel that anti racist communities is a better answer to this than tech solutionism. There needs to be a human step where an instance moderator makes decisions based on looking at evidence. Automating evidence collection and so forth is fine, but shared block list citation is not.

I'm white so I don't face racist harassment, but I was on Will Wheaton's block list for being trans and I don't want to replicate that kind of experience for others.

@celesteh @vfrmedia @scanlime @ariadne @timnitGebru fully agree with human-in-the-loop always, but surely there is a Grand Canyon sized chasm between

tech solutionism

and

reading through Federated and hope you see abuse before your users; manually reading other-instance.tld/about/more hoping they publish moderation, manually checking each instance, squinting at asterisks to see if they’ve been blocked before, and manually importing them one by one?