AI is creeping into therapy work in a way that feels helpful and risky at the same time. I’m talking about what happens when we start using AI to write our reports, our notes, and even our thinking, especially when we’re overloaded and just trying to get through the week. I love AI for the right jobs, but I’m seeing a pattern: when the cognitive load is high, we reach for shortcuts, and the quality of our clinical reasoning can quietly slide.

We dig into why AI can’t truly formulate a clinical opinion, and why that matters when you’re writing for real humans and for funding bodies like the NDIS. A report needs a clear line from functional impairment to disability to recommendation. If the output is generic, vague, or just “words on a page”, it can create more work, not less, and it can leave you unable to stand behind what’s written. I also cover privacy and professional integrity, including the very real issue of AI generating fake references and the need to keep client information protected.

Then we get practical. I share where AI is genuinely useful for therapists: structuring case notes from your own voice memos, holding information outside your working memory, helping with content ideas, transcription, and even supporting scheduling systems when you set them up intentionally. I also talk about the specific risk for early career therapists, because using AI before you learn to think can delay the exact skills you need to grow into a confident clinician.

If this hits home, subscribe, share it with a therapist friend, and leave a review so more clinicians can find it. What’s one task you think AI should never touch in therapy practice?

Sign up to my newsletter.

Follow me on Instagram.
Connect on LinkedIn.