Not really. AI will be ubiquitous of course, but humans who will offer advice (friends, strangers, therapists) will always be a thing. Nobody is forcing this guy to type his problems into ChatGPT.
We are already seeing AI-reliant high schoolers unable to reason, who's to say they'll still be able to empathize in the future?
Also, with the persistent lack of psychiatric services, I guarantee at some point in the future AI models will be used to (at least) triage medical mental health issues.