Email or username:

Password:

Forgot your password?
Top-level
Captain Jack Sparrow

@bart

Is it possible to detect and block AI generated images?

2 comments
Bart Decrem :mammoth_gold:

@Captain_Jack_Sparrow I love that idea. Should be a core aspiration or value of the project- a place for humans, not bots and ai

mirabilos

@Captain_Jack_Sparrow @bart not without a nōn-insignificant amount of false positives, and not without hassle (picking tools, multiple tools, tools change over time, tools are mostly “AI” themselves, etc.)

If it’s got a mod team and the flood is not too intense, not doing it automatedly might suffice.

Otherwise, at least provide ways for people to contest the decision of such bots…

OT for this subthread, but the posts from moth.social don’t load on my instance for some reason: consider that federating with AP will mean people on other softwares (Pixelfeed, but also Akkoma/GotoSocial/Mastdarm, etc.) will want to interact with that, so make sure that works as well as it can. I’d personally be reluctant to register “just another account” on an instance of your new system since I have a perfectly working AP-federated GtS instance here myself, for example.

@Captain_Jack_Sparrow @bart not without a nōn-insignificant amount of false positives, and not without hassle (picking tools, multiple tools, tools change over time, tools are mostly “AI” themselves, etc.)

If it’s got a mod team and the flood is not too intense, not doing it automatedly might suffice.

Go Up