ANON_TOKYO said in forum #421566:
You need to mention what makes it suspicious in the flag reason, just saying it is tells approvers nothing. The wiki also mentions that AI models cannot generate resolutions not divisible by 8, not that divisibility by 8 is any sort of evidence for anything (and even if the resolution matches the examples given in the wiki there's still cases where the artist knowingly or unknowingly uses an AI image as a reference, which might cause the resolution to match).
Also just check width:1448 height:2048. 2048x is a dimension limit of Twitter mobile, and it's a power of two, both of which contribute to it being extremely common.
Are the first two reasons insufficient? Or was I unclear in what I meant, namely that their style fits typical AI-generated works and the rate in which they post their works on Twitter seems unnaturally high.
I see the resolution is not suspicious now but that was more of an "adding to the list of points of concern". I wasn't sure how specific I had to be in the flag reason vs how much would be implicit.
Updated by luntoer

