Live data from Hacker News

Nightshade: University of Chicago's attempt at poisoning AI art generators

arstechnica.com

1–3 of 3 posts

Re: Nightshade: University of Chicago's attempt at poisoning AI art generators

#3

I wish I was smart enough to understand how things like this work.

a human built language model machine intended to recreate human expressions and interpretations is easily confounded by intentional human subterfuge until some other human builds in a lie detector