The subsection requires that the person "knowingly materially misrepresent [...] that material or activity is infringing". Proving that the misrepresentation was made knowingly can be extremely difficult.
Yeah how are we ever gonna possibly prove that against an AI-powered tool. It's the same with why police so readily embraced bad AI tech. It's all about shifting the liability and diffusing individual responsibility
