Is there actually any good evidence of that?

Did you skip the first sentence or something?

The survey shows a strong correlation between those who dislike AI and those who have negative emotions (exhausted, disillisioned, etc).

You could argue the survey isn't good evidence if you want, but what you quoted was a conclusion made from the survey data.

Regardless, it's pretty obvious that a polarizing topic would have a negative effect on many people's mental health, and few things are more polarizing right now than AI.