Trust and Responsible AI

Is it safe to put client information into AI?

It depends on your industry, your client's approval and the type of AI you use. Some client information can go into AI, such as recorded meetings you have it summarize. In medical and mental health work, be extremely careful with HIPAA. Look into your software, get approval, and set AI guardrails first.

Is it safe to put client information into AI?

On this page

  • What does it depend on?
  • Can client information go into AI at all?
  • What should you do before you put client information into AI?
  • What if the AI changes?
  • What are AI guardrails?
  • Where can you get help setting yours?

Is it safe to put client information into AI? It depends on your industry, your client's approval and the type of AI you use. Some client information can go into AI. Some needs extreme care.

AI guardrails are the rules you set for where and how you use AI, what information it gets and what it does not. You are the human and it is the machine, so you set them.

This is a big question, and I treat it as one. It is one of the first things I work through with clients when we build a second brain, which I describe in Where should a small business start with AI? This post is general education, not legal advice. Mapphouse is independent and not affiliated with Anthropic.

What does it depend on?

Three things: the industry you are in, your client's approval, and the type of AI you are using.

Your industry

If you work in medical, therapeutic or mental health settings, you have to be extremely careful with HIPAA rules about what you do with client information. Some software systems with built-in AI are made to help, for example by helping a doctor write summaries. Look into what a system does before you rely on it.

Your client's approval

Get your client's approval before their information goes into AI.

The type of AI you use

Look into the AI platform you are using and how you are actually using the client information with it.

Can client information go into AI at all?

Yes. There is client information you can put into AI. If you are a coach, a consultant or a creative director, one strong use is recording meetings and having AI summarize them, so you can pull out direct information.

There are many powerful ways to use AI with client information to streamline your work. The point is to do it on purpose.

What should you do before you put client information into AI?

  1. Look into the software you are using. Look at the AI platforms for real.
  2. Be clear on how you are actually using the client information.
  3. Get the client's approval.
  4. If you change how you use their information, update them. Keep that communication open.

What if the AI changes?

AI changes. If you use Claude, for example, and its protocols or systems are updated, stay up to date. Understand how an update affects the information you put in, for whatever purpose you use it.

What are AI guardrails?

This goes beyond client information. It covers your own personal information too. You are the human and it is the machine, so you set the guardrails.

Ask yourself:

  • Where and how am I using AI?
  • What information does AI get, and what does it not get?
  • What do I, as the human, still want to control and have my touch points on?
  • What am I willing, and wanting, to hand over?

This is one of the first things we talk about when I set up a second brain through the MAPPed process.

Where can you get help setting yours?

If you would like to talk through your guardrails, book a free discovery call.

Key takeaways

  • Whether it is safe depends on your industry, your client's approval and the type of AI you use.
  • In medical, therapeutic and mental health work, be extremely careful with HIPAA.
  • Some client information can go into AI, such as recorded meetings you have it summarize.
  • Look into your software, get the client's approval and update them if your use changes.
  • AI changes, so stay current. Set AI guardrails for what it gets, what it does not and what stays yours.

Common questions

Can I put client information into AI?

Some of it, yes. It depends on your industry, your client's approval and the type of AI you use. Look into your software, get approval and be clear on how you use the information.

What about HIPAA?

If you work in medical, therapeutic or mental health settings, be extremely careful with HIPAA rules about client information.

What is a good use of AI with client information?

Recording meetings and having AI summarize them, so you can pull out direct information. This is an example for coaches, consultants and creative directors.

Do I need to tell clients if my use changes?

Yes. If you change how you use their information, update them and keep the communication open.

What are AI guardrails?

The rules you set for where and how you use AI, what information it gets and does not get, what you keep control of and what you hand over.

Get Reverence in your inbox

From my journal to your inbox. Antidotes on raising a tiny human, life in the mountains and building a business with intention. Sent twice a month.

Join Reverence

Keep reading

Here’s how we begin.

Let's talk about what your business needs today.

I can't wait to say "hi!"
Book a free discovery call

See how MAPPed works.

A second brain, foundational Claude setup and tools to make your work sustainable.
See MAPPed