If you’re in a direct care role, your clients are probably already telling you that they're experimenting with AI tools (such as ChatGPT) for work, school, or even emotional support between sessions.

You personally may have considered integrating AI as part of your documentation process. You probably already use some earlier forms of "AI" if you use speech-to-text, predictive text, or certain data analysis tools.

And if you are a particularly forward-thinking individual, you might have been thinking about the role of AI in our profession and the long-term impact of AI on our profession.

I’ll happily admit, I’ve played with these tools for non-clinical tasks, and they can be impressive. (These emails are all me though, I write for fun.)

I wouldn’t say that I find AI efficient, per se. There’s still a fair amount work to train the chat-bots.

They don’t readily “understand” the types of questions you’re asking or the types of answers you’re seeking, and that becomes more true as your questions become more specific.

Chat bots and generative AI can give the same kind of generic mental health advice you’d find on any responsible website, but they struggle with nuance.

They don’t really understand the things that happen beyond “talk to a licensed professional to learn more.”

I suspect, given the research requirements of our degrees, we’d be better at finding our own answers faster, for now.

Generative AI (Gen-AI) is a technical term to describe the kinds of AI that can pretend to chit-chat with you by picking words in the order most likely to make sense for the conversation. It’s currently relatively cheap for the user (but expensive for the companies and the planet), easy to access at any time, day or night, and creates the illusion of shallow human understanding.

I can see the appeal. I think of it as a good overview of the content on the internet.

It can assume what words should go in what order. That’s helpful if you’re feeling jumbled up or brain-foggy when trying to write.

Gen AI can give you a good first draft if you’re struggling with the “blinking cursor problem”…as long as you’re a rigorous fact checker.

Several companies have emailed me to offer their (generative) AI notes tools for my practice. And I find myself…annoyed, I guess?

It’s concerning that AI tools are being developed and launched faster than we can expand our understanding, ethical guidelines, or regulations to keep us and our clients safe.

Would I love to not think about my notes anymore? Sure!

Do I trust an investment banker to start and run a company that will a) listen to my sessions, b) correctly identify the interventions, c) keep the individual data safe, AND d) use the aggregate data to make society better? Hmm let me think. OF COURSE NOT!

I’m not at the point yet where I feel able to make a recommendation about your workflow. There will undoubtedly be a useful role for some algorithmic technological support for therapists’ jobs.

If you’re an “early adopter,” someone who likes to be on the cutting edge of new tools and treatments...

...here are some of our current professional associations’ guidelines, recommendations, and intentions for the integration of AI into our field:

AI and Social Work (NASW)

https://www.socialworkers.org/About/Ethics/AI-and-Social-Work

The APA: “APA urges the Federal Trade Commission to put firm safeguards in place to prevent the public from harm”

https://www.apaservices.org/practice/business/technology/artificial-intelligence-chatbots-therapists

The Ethical Implications of Generative Artificial Intelligence for Psychologists

https://www.youtube.com/watch?v=IGXCyTyYS6k

Notice that each association is advocating for openness, understanding, and research, but many are also discussing the concerns of over-reliance on AI to assess, diagnose, or prevent self-harm.

Now MFTs, walk with me. Here we have:

CAMFT’s AI Ethics Recommendations

https://www.camft.org/Members-Only/Association-Documents/2025-2027-StrategicPlan/Technology-Integration

Hahahahahahahaha just kidding… CAMFT is planning to update our ethics code by 2027.

Let’s see what the AAMFT has for us…

Oh! Literally nothing related to AI is on the AAMFT website at all! Cool. (If you have seen otherwise, please let me know.)

So, for now, I’m going to be working with the ACA’s guidelines as my professional framework for AI in my practice.

ACA’s Recommendations for Clients:

https://www.counseling.org/resources/research-reports/artificial-intelligence-counseling/recommendations-for-client-use-and-caution-of-artificial-intelligence

ACA’s Recommendations for Counselors:

https://www.counseling.org/resources/research-reports/artificial-intelligence-counseling/recommendations-for-client-use-and-caution-of-artificial-intelligence

If you decide to utilize AI for your clinical practice, make sure you:

1. read ALL of the fine print: you’re looking for information about how data is stored, transmitted, owned, used, and deleted if necessary, as well as who is liable for what.

2. review your professional associations,

3. review your state guidelines,

4. get the approval of your leaders,

5. get informed consent from your clients,

6. and sign a Business Associate Agreement (BAA) with any company you’re planning to work with.

You may want to find a lawyer to help you understand any legalese. And remember that you are still working at the leading edge of this technology, so you may run into issues no one has anticipated yet.

There are still plenty of therapists who are moving cautiously with this technology, and others who are jumping right in!

Let me know if you’ve found a tool that’s made a big difference for you!

Chat with you next week!

-Andreana