There isn't a way to talk with AI zelots in a positive way.

So in your mind there is no middle ground people who feel somewhere between "AI boosting zealot" and "fuck AI it's terrible in literally every single way"?

If you honestly believe that then you've already spent too much time with simple minded communities that only deal in absolutes.

AI may not be a totally useless tool, but any implementation so far has been detrimental to human societies in every measurable way.

Sweeping broad statements like this are almost always incredibly easy to disprove. If you're talking about AI technology in general, of which LLMs are just one facet, then you can look at the protein structure work of Deep Mind which has literally revolutionized biomedical engineering by predicting the protein structures for hundreds of thousands of proteins, whereas previously predicting the structure of a single one would take a full research PhD.

Even if you're just talking LLMs there are a myriad of benefits from them being integrated into assistive technology, note taking for medical professionals, improved translation and transcription engines etc.

You can look for a community of ostriches who yell 'nah nah nah' while burying their heads in the sand, but you're not making a difference, you're just a bunch of birds ineffectively yelling at each other while blinding themselves to reality.