A Therapist’s Guide to AI: Ethics, Use Cases, and Simple How-Tos

Using AI safely and appropriately as a clinician

The need for ethical use of AI cannot be understated, especially in healthcare, but you know that already. Your clients’ well-being, safety, and privacy are all at stake. Whether you’re new to AI or already an active user, it’s important to stay up-to-date on how AI is changing, and how those changes bring both benefits and risks.

How AI reduces administrative burden for therapists

When you work in a deeply human profession like psychotherapy, it’s appropriate to be skeptical about how AI can fit into your day-to-day life. AI cannot replace the therapeutic relationship between a therapist and a client. What AI is proving appropriately useful for is reducing administrative workload.

According to a 2025 Alma survey, therapists who use AI tools do so because they save time (92%) and allow them to deliver better care (59%). In fact, 45% of providers who use AI tools agreed that more than half of their admin time could realistically be managed by AI without compromising quality of care.

In the same survey, the therapists who reported feeling most hopeful about the potential of AI to support mental health care providers were those who were already using AI tools.

AI tools can help you:

Write compliant progress notes—Whether you want to try automated note-taking or double-check your own notes to improve compliance, we’ll show you both approaches.

Improve your consultation strategy—By using AI in a roleplay capacity, you can surface gaps and opportunities in how you communicate with potential clients during consultations.

Make effective marketing materials—Smart prompts can assist you in creating a compelling provider profile, brainstorming social media posts, and fine-tuning your blog strategy.

Comparing HIPAA-compliant AI tools for therapists

If you’re serious about making AI a mainstay in your practice, consider investing in a HIPAA-compliant GPT. These AI platforms specialize in data security and are designed to protect your clients’ protected health information (PHI).

With HIPAA compliant GPTs, you can securely upload a full progress note, including PHI, and prompt the GPT to review it against a list of compliance requirements.

Examples of compliant GPTs include:

• CompliantChatGPT: This AI agent tokenizes PHI to ensure patient privacy and confidentiality. You can use it to create treatment plans, analyze patient cases, get clinical notes created from uploaded text or audio, and more. It uses the Gemini 2.5 Flash Lite AI model.
  • The starter plan starts at $19.99 per month.
• BastionGPT: This AI tool can help you take session notes, create comprehensive evaluation reports, and generate ideas to adjust treatment. It can also help with admin tasks like follow-up emails after sessions and appointment reminders. It incorporates specifically trained versions of multiple AI models (ChatGPT, Claude, and Gemini).
  • Plans start at $20 monthly.
• Hathr: This model stores all information on AWS GovCloud, which ensures the utmost security. Hathr can convert uploaded notes into custom formats, help you identify diagnostic patterns, and automate admin processes like insurance verification, intake, and appointment reminders. It uses the Claude 4.0 model.
  • Plans start at $45 monthly.

Is your clients’ PHI protected? The details of how PHI is protected by each HIPAA-compliant GPT will be outlined in their Business Associate Agreement (BAA). This agreement specifies their security measures, how they’ll notify you of any issues, and that patient information can only be used for the specific services they’re providing to you. When you choose an established, reputable HIPAA-compliant GPT platform and ensure they sign comprehensive BAAs, you’re building multiple layers of protection around your patients’ information.