Using AI Ethically

Artificial intelligence is becoming a powerful tool for therapists, offering practical support in documentation, workflow automation, and even client engagement. But the big question remains: how can you use AI ethically without putting client confidentiality at risk? As a therapist committed to HIPAA compliance and ethical standards, this post will walk you through how to use AI responsibly while protecting client trust every step of the way.

This guide is especially for mental health professionals like LPCs, LCSWs, psychologists, and psychiatric nurse practitioners who want to leverage AI in their practices without crossing ethical or legal boundaries.

What Counts as Client Confidentiality?

Confidentiality is the ethical obligation to protect private information shared by clients in a therapeutic relationship. It’s central to what we do, and essential to maintaining client trust. For therapists in the U.S., confidentiality isn’t just a professional code, it’s also backed by federal law via HIPAA.

HIPAA’s Privacy Rule protects any client information that can be used to identify a person and relates to their physical or mental health — also called Protected Health Information (PHI). That includes everything from names and dates of birth to session notes and treatment plans.

So, anytime you’re using a digital tool that touches PHI, you need to make sure that tool is secure and compliant.

Where AI Fits Into the Picture

AI tools can be incredibly helpful — especially when it comes to note-taking, treatment planning, summarizing sessions, or organizing documentation. But with that power comes the need for extra caution.

Here’s what it means to use AI responsibly in therapy:

• You don’t upload client info into tools that aren’t specifically built for HIPAA compliance
• You keep all PHI secure — both in transit and at rest
• You have clear consent from clients when using AI tools during or after sessions
• You always review AI-generated content for accuracy before saving it to a clinical file

Key Guidelines for Ethical AI Use in Therapy

Here’s a breakdown of best practices that ensure your use of AI protects confidentiality, maintains ethical standards, and keeps you compliant:

1. Only Use HIPAA-Compliant AI Tools

If you’re using an AI assistant or note-writing tool in your practice, it must be HIPAA-compliant. This means:

• The company will sign a Business Associate Agreement (BAA)
• Data is encrypted in transit and at rest
• Access to PHI is restricted and secured

One tool that checks all those boxes is HeyBerries.com/therapists. As a clinical consultant for Berries, I can confidently say it’s built specifically for mental health professionals and is structured around ethical and secure use of AI in private practice.

Avoid tools like general-purpose AI assistants that aren’t designed for medical data — even if they seem helpful. If they’re not HIPAA-compliant and you upload client info, it could result in a breach.

2. Use Minimal and De-Identified Information

When working with AI, follow the rule of minimum necessary data. Don’t upload anything you don’t absolutely need to — and anonymize when possible.

If you’re using AI to summarize sessions or generate notes, tools like Berries allow you to work from audio while still giving you control over what data is stored, seen, or shared. No client-identifiable data is retained or used to train the AI model.

This ensures that even if something goes wrong, no traceable PHI is floating around in the system.

3. Get Informed Consent from Clients

Transparency is everything. If you’re using AI tools — especially ones that involve transcription or summarizing sessions — your clients should know. Always:

• Explain why and how the tool is being used
• Share what data (if any) is collected or stored
• Reassure them that the tool is HIPAA-compliant
• Allow clients to opt out if they’re not comfortable

You can include a tech-specific consent form in your intake paperwork or have a brief conversation about it when introducing new tools into your workflow.

Example script: “I use a secure AI assistant called Berries to help with documentation so I can stay more focused during sessions. It processes audio from our session to help me create a clinical note. Everything is private, secure, and meets HIPAA standards.”

4. Keep the Human in the Loop

AI should assist, not replace, your clinical judgment. That means you — the licensed professional — are always the final reviewer and editor of anything AI-generated.

Don’t copy and paste blindly. Review every note, ensure it reflects the session accurately, and make edits where necessary.

If the AI inserts something off or clinically incorrect, you’re responsible for catching and correcting it. That’s true whether you’re writing a progress note, a treatment plan, or a clinical letter.

5. Make Security a Habit

Even with HIPAA-compliant tools, your end of the process matters too. Here’s how to keep things secure on your side:

• Use strong, unique passwords
• Enable two-factor authentication where available
• Don’t leave client notes or AI tools open on shared devices
• Keep your devices updated with the latest software

Also, be aware of your environment. If you’re dictating a session summary into your AI assistant, make sure you’re in a private space, just like you would be during a therapy session.

6. Delete What You Don’t Need

If the AI tool you’re using stores transcripts or session audio, know what their data retention policies are — and have control over when and how to delete that data.

With HeyBerries, for example, session audio is automatically deleted after the note is created. You can also delete any content inside the app at any time. This limits liability and keeps your practice more secure.

7. Stay Informed and Evolve With the Guidelines

The ethics of AI in mental health are evolving. State boards, national associations, and licensing bodies are beginning to release more specific guidelines.

Make it a point to:

• Follow updates from APA, NASW, ACA, and your specific licensing board
• Attend CE events or webinars on ethical tech use
• Join communities like The Clinical AI Club where we discuss these developments and help therapists adapt

Final Thoughts: You Can Use AI Ethically and Still Sleep at Night

AI doesn’t have to be scary — or unethical. If you’re intentional about how you use it, stick with tools built for our profession, and keep client privacy front and center, it can actually make your practice more ethical.

Less burnout. More focus in session. Better records. Happier clients.

We just have to do it right.

If you’re curious about what that looks like, come check out www.clinicalaiclub.com. We talk about this all the time — the wins, the missteps, and the real-life impact of using AI in therapy the right way.

And if you’re ready to try a HIPAA-compliant AI tool that truly supports your practice, visit HeyBerries.com/therapists. It’s built by therapists, for therapists — and you’ll see the difference the first time you use it.

You don’t have to choose between efficiency and ethics.

You can have both.