“Hiding hate symbols within innocent images has become a way to spread hate across the web, so much so that the practice arguably forms its own strand of generative AI design,” the report stated.