Enhancing therapy with AI: A guide for clinicians and clients

As technology continues to evolve, finding ways to integrate it into our therapeutic practices can be both exciting and beneficial. With the launch of ChatGPT-4 earlier this year, we have even more opportunities to enhance therapy with AI. At the same time, it's crucial to be mindful of how we're integrating this technology, just as we would with any other new tool or practice. I hope this article can be a useful guide for clinicians and clients who might be considering, or already using, AI tools in their work together.

With any free or new technology, I have learned to ask the question “what’s the cost?”. Often with tech, it’s our data. In the case of ChatGPT and other AI, I was dismayed to learn that the answer to “what’s the cost” is also the environment. For those who haven’t heard, training and utilizing AI uses a shocking amount of energy, leading to increased electricity usage and carbon emissions: it's estimated that "the carbon footprint of training a single big language model is equal to around 300,000 kg of carbon dioxide emissions. This is of the order of 125 round-trip flights between New York and Beijing". It also uses an excessive amount of water: while "a single Google search requires half a millilitre of water in energy, ChatGPT consumes 500 millilitres of water for every five to 50 prompts". There are even more environmental and human costs, and companies are not being fully transparent about them.

I don’t know what the answer is here, but I think it’s important to start with awareness of the costs and benefits of using certain technologies, especially for the population groups that are likely to suffer most from climate change. I want to be clear that I’m not here to say whether an individual person should use AI or not - especially because I have seen some incredibly positive and helpful uses despite being wary about the costs. Instead, I thought it would be helpful to add to what I hope will be an ongoing conversation about the shifts we’re making as a society, and if they are truly in line with the future we’re trying to co-create.

What is the most ethical answer?

The thing is, personally, I’m still torn. But in my work as a therapist,  I’m familiar with having two important values that come into conflict and having to navigate this.  What’s good for the environment may come into conflict with what’s helpful for someone’s mental health in many ways. For example, I have often encouraged clients to get their groceries delivered, order take out, or do whatever is necessary to make sure that food actually goes into their bodies, even when it’s not the most financially or environmentally sustainable option. I have also encouraged people to take a cab or separate car instead of carpool, when doing so would mean they can actually attend a social function and not worry about their safety or ability to leave when their nervous system said it was time. I have also supported people to use AI to maintain their ability to keep working in the job they’re in, or to help them attain new work that doesn't burn them out.

These are small examples of how everyday we make choices where we have to compromise on our values, and decide out of all the things that are important to us, what’s most important. This has to be an evolving and conscientious process. I have seen often that people get stuck when we create rigid rules around what is considered “ethical” and what is not, and then place those rules onto other individual people who are struggling. 

What I’ve also come to recognize is that putting the burden of the environment onto the shoulders of individuals is unfair when it is primarily large corporations that are causing the current climate crisis, and it also confounds the real problem of our capitalist system. This burden of ethical consumerism also disproportionately impacts the people who need those easier, cheaper options (e.g., disabled folks, people living below the poverty line, people with limited time and energy).

Important considerations of using AI in therapy

All that preamble aside, incorporating AI into our therapeutic practices requires thoughtful consideration. Here are some key questions and points to keep in mind: 

• Data privacy and security: Where and how is the information being stored and used? Ensuring that client data remains confidential and secure is paramount. 
• Accuracy and reliability: Am I double-checking the accuracy of the information provided? It's important that AI supplements, rather than replaces, our professional judgment and expertise.
  
• Continuous education: Engage in ongoing education about the use of technology in therapy. Understand the capabilities, limitations, and ethical implications of using AI tools like ChatGPT in your practice.
  

Additional ethical considerations

• Informed consent: Ensure that clients are fully informed about how AI is being used in their therapy. Explain the potential benefits and limitations, and obtain their explicit consent.
  
• Transparency: Be open with clients about the use of AI tools. Explain how AI generates responses and the nature of its algorithms to ensure clients understand it’s an aid, not a substitute for human expertise.
  
• Bias and fairness: AI models can reflect and perpetuate biases present in the data they were trained on. Be mindful of potential biases in AI’s responses and critically evaluate the information it provides.
  
• Scope of use: Clearly define the scope of AI’s use in therapy. Again, it should enhance, not replace, professional judgment. Ensure all critical therapeutic decisions are made by the therapist, not the AI.
  

Possible positive uses of AI in therapy

Here are some positive ways in which both clinicians and clients can potentially take advantage of AI:

Positive uses for clinicians

• Personalized guided visualizations or meditations: Create customized visualizations or meditations tailored to your clients' needs. For example, prompt AI with “create a guided visualization to help with [THEME: anxiety, compassion], using imagery of [SETTING: forest, pathway].”
  
• Worksheet and journal prompt creation: Generate worksheets or journal prompts to help clients explore their thoughts and feelings in a structured way.
• Resource recommendations: Ask AI to suggest relevant articles, books, or therapeutic resources that might benefit your clients.

Positive Uses for Clients

• Project breakdown: If you're feeling overwhelmed by a project (e.g., cleaning your house), ask AI to break it down into manageable, bite-sized tasks.
  
• Assertive communication scripts: Get help formulating scripts for assertive communication, making it easier to express your needs and boundaries.
• Grounding strategies: Request a reminder of grounding techniques, such as deep breathing, mindfulness exercises, or sensory activities, to help manage stress and anxiety.

Final thoughts

As we've seen, AI isn't an evil but it's also not an ultimate solution. It's a human-made tool with potential benefits as well as concrete negative impacts we're only now beginning to uncover. I would highly recommend this briefing paper of CPA on AI and psychology as further reading for psychologists, as well as this article on the negative impacts of using AI in science research, to keep you thinking about this topic. In addition, this video by theoretical physicist Angela Collier is a good, entertaining, and easily digestible primer into what AI is and what it can't do.

How about you? I would love to hear if you're using AI tools in your practice, what you're using them for, and how you're talking with clients about it. .