I feel a bit cautious about the counsellor forecast because anecdotally and in surveys, therapy seems to be one of the most popular uses of AI at the minute.
Mostly this seems to be new demand – people are speaking to their AI therapist every day and then also speaking to a human therapist once a week, or are using AI therapy when they wouldn't have bothered to get a human therapist before. And I think people could do 10x more therapy than they do today.
So I don't expect an immediate impact on employment.
But AI therapy seems like seems likely to continue to improve as models get better memory of past interactions, plus better social skills as they get larger (e.g. GPT-4.5 vs 4), more personality (maybe realistic video avatars). In combination with advantages like unlimited free 24/7 availability, infinite patience and ability to provide any therapy style, it wouldn't be surprising if people stopped bothering to see their human therapist.
I'm aware of the popularity of AI therapists and would agree it seems to be mostly creating new demand-- either providing a lifeline to those who can't afford a human therapist or enhancing the experience of those who already see a one regularly.
There are a couple of reasons why I think the demand for therapists will be more durable. First, there's what @danablochtherapy calls "exquisite presence”:
There is some hard evidence of this in telehealth data-- mental health care peaked at 66% virtual at the height of the pandemic, but fell to ~37% by the end of 2023. So lots of people who might have had the option to see someone virtually decided that the trouble of going in-person was worth it. I think it will be difficult for AI to fully replace what people have shown this revealed preference for:
Maybe there will eventually be some erosion in demand for therapists from those who might otherwise prefer virtual appointments. And maybe overall mental health could improve due to other factors. But one final factor I am curious about is if the close connection between physical and mental/social health might eventually push the occupation to evolve in the direction of coaches and personal trainers. Occupational definitions are likely to continue to evolve as they have done in the past.
I feel a bit cautious about the counsellor forecast because anecdotally and in surveys, therapy seems to be one of the most popular uses of AI at the minute.
Mostly this seems to be new demand – people are speaking to their AI therapist every day and then also speaking to a human therapist once a week, or are using AI therapy when they wouldn't have bothered to get a human therapist before. And I think people could do 10x more therapy than they do today.
So I don't expect an immediate impact on employment.
But AI therapy seems like seems likely to continue to improve as models get better memory of past interactions, plus better social skills as they get larger (e.g. GPT-4.5 vs 4), more personality (maybe realistic video avatars). In combination with advantages like unlimited free 24/7 availability, infinite patience and ability to provide any therapy style, it wouldn't be surprising if people stopped bothering to see their human therapist.
I'm aware of the popularity of AI therapists and would agree it seems to be mostly creating new demand-- either providing a lifeline to those who can't afford a human therapist or enhancing the experience of those who already see a one regularly.
There are a couple of reasons why I think the demand for therapists will be more durable. First, there's what @danablochtherapy calls "exquisite presence”:
https://substack.com/home/post/p-160985866
There is some hard evidence of this in telehealth data-- mental health care peaked at 66% virtual at the height of the pandemic, but fell to ~37% by the end of 2023. So lots of people who might have had the option to see someone virtually decided that the trouble of going in-person was worth it. I think it will be difficult for AI to fully replace what people have shown this revealed preference for:
https://www.epicresearch.org/articles/telehealth-utilization-higher-than-pre-pandemic-levels-but-down-from-pandemic-highs
Maybe there will eventually be some erosion in demand for therapists from those who might otherwise prefer virtual appointments. And maybe overall mental health could improve due to other factors. But one final factor I am curious about is if the close connection between physical and mental/social health might eventually push the occupation to evolve in the direction of coaches and personal trainers. Occupational definitions are likely to continue to evolve as they have done in the past.