I’m not putting all my negative feelings into a company sponsored LLM to fucking have it say, “no relax guy, it’ll be OK.” Like it’s a fucking clanker.

I'd be more concerned with any company sponsored AI chatbot therapist using what you say influence your employment relationship.

Employee X: I'm worried about losing my job so I work unpaid overtime and that is affecting my marriage.

Therapist chatbot to management: Employee X should not be given a raise. They already have enough external motivation to work without additional financial incentives.

Well, I'm sure someone who uses "clanker" wouldn't need therapy anyway.

Seriously though, I doubt the health implications or claims about the efficacy of AI therapists, but we can't just ignore the fact that there are people who use it, which means there's something about it that makes it accessible or preferable to a human therapist.

If you've ever had to get a psychotherapist, you know that it is prohibitively expensive for a large number of people, and that a human therapist may not actually be capable of treating you because of personal incompatibility; which often results in retraumatization in patients who are seeking therapy for particularly traumatic or sensitive issues. Since much of the value in therapy is learning management strategies that, while not standard, are often consistent across different practitioners, they do not necessarily need to come from a therapist to learn what they are (even if the practice of them does need one).

I think if there is a need for it, that need is a consequence of the deeply dysfunctional, exploitative, and isolating system we live under, and I don't think I'd ever accept it as a genuine alternative to human therapists. But, we can't dismiss it out of hand if there are people who say it is useful for them and when we can't maintain a system that can guarantee them access to treatment.