How AI Tools Can Help Therapists, According to Therapists

Here’s everything you need to know about how therapists might use AI tools ethically.

Learn more about our Clinical Review Process

Artificial intelligence is infiltrating so many parts of life, from work to medical appointments. Increasingly, some clinicians are using AI tools within their work. No, that doesn’t mean they’re outsourcing mental health support to an AI therapist; it may mean that they’re using AI tools within their traditional therapy practice.

So, what are the best AI tools for therapists, and how should they be used? “The safest approach is to use AI as a support tool — not as a clinical replacement,” says Charlie Health Group Facilitator Bree Williams, LPCA. “Clinicians should choose platforms specifically designed for behavioral health that are HIPAA-compliant, encrypted, and transparent about data usage.” Below, we dive further into ethical AI tools for therapists, the potential risks, and best practices.

How many providers use AI tools?

According to the American Psychological Association’s (APA) 2025 Practitioner Pulse Survey, over half (56%) of psychologists used AI at work in any capacity — up from 29% in 2024. However, the number of psychologists with worries about AI also grew. At least 60% of psychologists said they were concerned about: 

• Potential data breaches 
• Unanticipated social harms 
• Biased training data leading to biased outputs
• Lack of rigorous testing to mitigate risks
• Inaccurate outputs and hallucinations 

What AI tools are available for therapists? 

Many clinicians use AI tools based on their own research. Charlie Health Licensed Creative Arts Therapist Courtney Way, MA, LCAT, for instance, said she turns to widely known AI models or tools. Meanwhile, Williams points to tools like Mentalyc, Upheal, Blueprint AI, and AutoNotes, which she says are examples of “platforms designed for behavioral health workflows.”

Both clinicians emphasize that whichever AI tools are used should meet certain criteria — and only be used for specific things. Put simply: AI tools should be focused on administrative tasks, like note-taking, scheduling, and progress summaries. In some cases, these tools might also help with symptom tracking and treatment planning. They aren’t a substitute for critical thinking or therapy-client interactions. 

Are there ethical AI tools for therapists? 

Many clinicians use AI tools based on their own research. Especially since clinical ethics codes haven’t necessarily caught up to AI use. According to Way and Williams, some of the most ethical tools focus on secure data handling, allow for significant oversight, and have HIPAA compliance. “It is very important to look out for systems that will meet our privacy and confidentiality guidelines,” says Way.  

Also, the ethics of using AI tools in clinical settings may depend on how they’re used, the experts said. “I think ethical AI in therapy works best when it reduces administrative burden — not when it attempts to replace clinical judgment or human connection,” says Williams. “When used responsibly, these tools can give therapists more time and emotional bandwidth to focus on actual client care instead of spending hours completing documentation after sessions.”

What risks can AI tools pose for therapists and clients? 

Clinician use of AI might help offload administrative burdens, but it can also bring a number of challenges and concerns for both the therapist and their clients. “The biggest concerns involve privacy, overreliance, accuracy, and ethical responsibility,” says Williams. 

1. Privacy concerns 

It’s absolutely critical that clients are made aware if, how, and, why a clinician is using AI, says Williams. Everyone deserves to know how their information is being processed. 

Clinicians must pay attention to how a person’s mental health information is stored and if the AI system is using that data for training. “Some general-purpose AI tools are not built for healthcare compliance, which creates serious confidentiality risks,” says Williams. 

2. Overreliance on AI

Where is the line between AI being helpful and AI being overused? Way notes that clinicians run the risk of relying too much on AI, something that might harm their judgment or the therapist-client relationship.

AI isn’t a replacement for human understanding, and an overreliance on such tools can harm people. “AI can summarize information, but it cannot truly understand nuance, attachment dynamics, trauma history, or nonverbal cues in the way a trained clinician can,” Williams explains. “Researchers and clinicians have also raised concerns that AI systems may miss suicidal ideation, overvalidate harmful thinking, or create a false sense of emotional understanding.” 

Certified human mental health professionals need to be the ones making a treatment plan, looking at client progress, and running therapy sessions. 

Best practices for clinicians using AI tools 

When traditional therapy — whether in-person or online — brings AI into the mix, there are necessary steps to ensure excellent patient care and safety. 

1. Double-check everything

Again, there’s no room for overreliance on AI. A human therapist must make the decisions and review everything after each session. Maintaining full oversight means looking at every AI-generated note, treatment suggestion, and summary with a critical eye, says Williams. A clinician needs to review and edit session notes and clinical documentation before they become a part of a patient’s record.

2. Maintain data protection 

Data protection is critical when using AI with clients. “It’s important to make sure that what is documented remains protective, with enough information to know what happened in session, while still maintaining the protection of our clients,” says Way. Choose tools that offer clear security and will remain HIPAA-compliant. 

Williams seconds this: “Therapists should avoid entering unnecessary identifying details into unsecured systems and should clearly communicate with clients about how AI is being incorporated into care.”

3. Keep the therapist-client relationship at the forefront 

These tools are not meant to be an AI therapist. The relationship between a human therapist and their client is a very important bond filled with emotional support. That connection is key to any mental health care intervention, says Williams. “Empathy, attunement, clinical judgment, and relational safety cannot be automated,” she adds. “AI should enhance a therapist’s ability to be present, not replace the human connection clients come to therapy for.” 

How Charlie Health can help 

If you or a loved one is struggling with your mental health, Charlie Health can help. Charlie Health is a virtual behavioral health provider delivering high-acuity treatment for kids, teens, and adults facing serious mental health conditions and substance use disorders. Our innovative treatment model combines clinical expertise, group connection, and measurement-based care to support long-term healing. With the right support, managing your mental health and relationships is possible. Fill out the form below to start healing today.ay.