Part 1 | Meeting generative AI
Chapter 1. Why therapists should learn generative AI now
Key points
- Treat generative AI as an aid to a therapist, not a replacement for one
- Therapists turn out to be an unusually good fit with AI
- What matters is less how much you know, and more how well you can ask
The feeling of being left behind
The phrase "generative AI" has become impossible to avoid. Since ChatGPT appeared in 2022, AI has spread through ordinary life, and many tools have arrived alongside it:
- ChatGPT
- Gemini
- Claude
And beyond them, new services keep being born:
- image generation
- video generation
- AI search
- AI coding
- AI agents
Honestly, it moves fast enough that plenty of people simply cannot tell what is going on.
In the therapy world in particular, the questions are anxious ones: can this be used in the clinic, will the work disappear, where do I start, is patient information safe? Meanwhile the information itself is scattered, and there are few places to learn it in an organised way.
Which is exactly why it is worth stopping once to lay out, as practically and realistically as possible, how a therapist might actually use generative AI.
AI is an aid, not a replacement
Start from the premise that, as things stand, AI is unlikely to replace a therapist outright.
A therapist is a professional who
- observes bodies
- understands context
- reads emotion
- builds relationships
- supports behaviour change
Non-verbal communication, bodily sensation, the feel of a room, trust, ethical judgement — these are precisely the areas where today's AI is weakest.
So it is more realistic to think of AI not as a stand-in for the therapist but as
an intellectual pair of training wheels
In the way training wheels hold up someone who cannot yet ride, AI quietly supports the knowledge you have not yet put into words and the information you have not yet organised.
Therapists and AI fit together better than you would think
There is a real sense in which this profession suits AI unusually well, because a therapist spends every day doing:
- organising information
- forming hypotheses
- explaining
- teaching
- learning
- documenting
- reasoning
which is very close to a list of what generative AI is good at.
For example, these are easy to hand over:
- searching papers and summarising them
- drafting explanations for patients
- outlining study-group material and case reports
- writing multiple-choice questions for national exams or in-house tests
- organising literature in several languages (English, Chinese and so on)
When you have to relearn a disease or a presentation you have never met before, AI makes a powerful first teacher.
The ability to move across PubMed and Google Scholar while organising material in several languages is a particular advantage for therapists who find foreign-language literature intimidating.
In the age of AI, knowing a lot is not enough
For a long time, how much you had memorised mattered a great deal. From the national exam onward, therapists have been asked to hold an enormous amount of knowledge in their heads.
From here on, though, what grows in importance is
how you organise knowledge, and how you use it
AI can handle volumes of information. That is exactly why more is now asked of the human side:
- the ability to structure
- the ability to ask
- the ability to read context
- the values that decide what matters
What this guide covers
We will work through the basics of ChatGPT, its use in clinical work and teaching, file uploads, prompt technique, and GPTs and Gems — a little at a time. Rather than memorising difficult terminology, the emphasis is on
"just try it"
What matters in this era is not understanding perfectly before you begin, but
updating your understanding while your hands are on it
Chapter 1 summary
- AI is not a replacement for a therapist; it is a set of intellectual training wheels
- A therapist's daily work — organising, reasoning, teaching — fits AI well
- The ability to ask good questions matters more than the volume you have memorised
- More than anything else: touch it