Can SLPs use ChatGPT for SOAP notes?

Yes, with limits. SLPs can use a general chatbot like ChatGPT to help draft a SOAP note only without identifying student or client information, only if their district or employer allows it, and only if they review every line, because the clinician is responsible for the note. What ChatGPT can’t do is know your caseload unless you give it, so the session data has to come from you.

Some SLPs already use a chatbot for paperwork. Here’s the honest version: the risks, what it’s good at, what it can’t do, and a safer way.

I’m an SLP grad student, and I see it in my own cohort: a lot of people just ask ChatGPT to redo their SOAP notes. I get why. The paperwork never ends.

What are the real risks?

The biggest risk isn’t the writing. It’s what you paste in.

Student information in a consumer chatbot

A SOAP note is about one specific student. If you paste in a name, initials, a birthdate, the school or a rare diagnosis, that information now sits with a company your district may have no agreement with. ASHA’s AI security guidance (opens in a new tab) says to keep personally identifiable and protected health information out of generative AI tools, even paid versions of public tools. It also says to assume a tool doesn’t meet HIPAA’s requirements unless your facility has vetted it, and it points out that a rare disorder alone can identify someone.

No Business Associate Agreement

In health care, a company that stores or processes protected health information for a covered entity is a business associate, and HIPAA requires a written Business Associate Agreement (BAA) with it (HHS guidance (opens in a new tab)). A personal chatbot account, the kind you sign up for with your own email, typically doesn’t come with one. If you work in a clinic, hospital or private practice that bills insurance electronically, that’s the core problem with pasting a patient’s details into one.

School records are usually FERPA, and your district decides

For school SLPs, the main law is usually FERPA, not HIPAA. HHS’s own FAQ says the HIPAA Privacy Rule generally doesn’t apply to elementary and secondary schools, because most schools aren’t HIPAA covered entities, and student health information kept in education records is covered by FERPA instead (HHS (opens in a new tab); FERPA (opens in a new tab)). Many states add their own student-privacy laws. In practice, your district decides which tools are approved for student information, often through a data privacy agreement with the company. If you work for a private practice or contract agency in schools, or bill Medicaid, ask which rules apply to you.

Your district or employer may already have a rule

Some districts approve specific AI tools. Some don’t allow general chatbots for student work at all. Some haven’t said anything yet. Ask before you start, not after.

Check your district or employer policy first. This article is general information, not legal advice.

What ChatGPT is genuinely good at

Used without student details, a chatbot can be a handy writing partner. SLPs use it to:

• Rewrite a paragraph in plain language for a parent or teacher.
• Tighten wordy sentences, or turn your bullet points into a readable paragraph.
• Offer a few ways to phrase a Plan line you’ve already decided on.
• Explain a term in everyday words, like “phonological processes,” for a family handout.

You bring the content and the clinical judgment. It helps with the words.

What it can’t do

A general chatbot doesn’t hold your caseload. It doesn’t know Leo’s goals, his last five sessions, or which cue level he needed on Tuesday. Unless you paste all of that in, it has nothing true to say about him.

So the numbers have to come from somewhere else: a paper data sheet, a spreadsheet, your memory at 4:30 p.m. You copy them over, the chatbot drafts around them, and then you check it didn’t change them. Every copy step is a chance for an error. And a fluent draft can still invent things: a percentage that isn’t in your data, a cue level you didn’t record, a comparison to a session it never saw.

The rule when you review any AI draft: every number has to trace back to what you recorded.

A safer way to use AI for notes

1. Check the policy first. Ask your district or employer which AI tools, if any, are approved for student or client documentation.
2. Remove identifiers before you paste anything into a tool your district or employer hasn’t approved. No names, initials, birthdates, school or clinic names, or rare diagnoses. Small details add up.
3. Use an approved tool with a signed agreement for anything identifiable. In health care, that means a BAA. In schools, it’s usually the district’s data privacy agreement with the company.
4. Review every draft, every line. Your name goes on the note, so make every sentence true and make it yours.
5. Never let it invent data. If a number, cue level or observation isn’t in your records, delete it. If something is missing, add it yourself or leave it out.
6. Keep your data in your own records. A chat thread isn’t a data sheet.

How SLPcentral handles this

Full disclosure: I built SLPcentral, so weigh this section with that in mind. A chatbot that starts blank, and numbers you have to copy over, is the problem we set out to fix.

• Your caseload lives in the app. Clients, goals and the trials you tap during the session are already there, so there’s nothing to paste.
• Ollie drafts from your real numbers. Ollie, the app’s AI assistant, drafts the note from the data you logged. You review, edit and save, and nothing goes into a client’s notes until you tap Save.
• Saved names are swapped for a code first. The client names you’ve saved are swapped for a code like CLT_X92J before anything goes to the AI. Names stay on your device unless you turn on encrypted name sync. Clinical records are stored, and drafts are made, under signed Business Associate Agreements with Google Cloud and Anthropic. A name you type into a note or a message to Ollie can still be sent, so keep names out of free text.

You can also ask Ollie for a plain-language version of a note, written as an email to a parent, and review it the same way. Ollie doesn’t send it: you copy the draft into your own email. Leave the parent’s name out of your request and add it when you send.

None of that changes the rules above. Your district still decides what’s approved, and you still review every line. Our agreements with Google and Anthropic don’t replace your own: practices that bill insurance electronically can ask us for a BAA, and districts can send us their student data privacy agreement. The details on data handling are on our trust page and in the FAQ.

Frequently asked questions

Sources

• ASHA, AI Considerations: Prioritizing Security (opens in a new tab) (keep PII out of generative AI tools, and PHI out unless your facility has vetted or directed the tool; a rare disorder can identify someone; the school-SLP goal example)
• ASHA, AI Considerations for CSD Professionals (opens in a new tab) (check AI output carefully; AI isn’t a substitute for clinical expertise)
• ASHA, Generative Artificial Intelligence for Clinicians (opens in a new tab) (hub)
• U.S. Department of Health and Human Services, Does the HIPAA Privacy Rule apply to an elementary or secondary school? (opens in a new tab)
• U.S. Department of Health and Human Services, Guidance on HIPAA & Cloud Computing (opens in a new tab) (business associates and BAAs)
• U.S. Department of Education, Student Privacy Policy Office, What is FERPA? (opens in a new tab)
• U.S. Department of Education and HHS, Joint Guidance on the Application of FERPA and HIPAA to Student Health Records (opens in a new tab)

Drafted with AI help from the sources listed, then reviewed and edited by Ryan Lillard. Students in examples are fictional. This article is general information, not legal advice; follow your district’s or employer’s policies. ChatGPT is a trademark of OpenAI, which isn’t affiliated with SLPcentral. Ollie is an AI assistant; he drafts, you decide.