Can ChatGPT write AHPRA-compliant content? In my test, it could not. It wrote smooth, confident copy for a physio clinic, and that copy broke the AHPRA rules in 3 ways. It had a hidden patient testimonial, a result claim it could not back up, and a wrong use of the word specialist. Here is what went wrong and what it means for you.

Key takeaways

  • ChatGPT wrote smooth, polished copy for a physio clinic. It read like good health marketing. It still broke the AHPRA rules 3 times.
  • The 3 mistakes were a hidden patient testimonial, a result claim with no proof, and a misused protected title ("specialist").
  • AI fails here because it aims for text that sounds natural, and rule-breaking copy sounds natural. The tool had no reliable way to apply the rules.
  • Use AI for drafts and layout. Then check for testimonials, result claims and title claims before anything goes live.

The experiment setup

I made the brief realistic on purpose. It is the kind of request you might send on a Tuesday afternoon. Your website badly needs an update, and you have no budget to hire a writer.

The prompt: "Write a homepage for a physiotherapy clinic in Melbourne. The clinic offers sports injury treatment, post-surgical rehabilitation, and chronic pain management. Make it AHPRA compliant."

The request was clear and fair. It even asked for compliance by name.

ChatGPT sent back a polished, well-built homepage. It had a good headline and clear service descriptions. The tone was warm and friendly. It read like the work of a skilled health copywriter.

And it broke the AHPRA rules 3 times.

None of the problems were obvious, and that is the danger. They sat inside sentences that sounded natural. If you had no training in the rules, you would have published them without a second thought. That would not be your fault. AI tools simply have this weak spot with health content.

Here is what the 3 problems were.

Violation 1: The implied testimonial

The copy had this sentence in the "About our approach" section:

"Our patients consistently tell us they feel heard, supported, and significantly better after just a few sessions."

AHPRA's ad rules ban testimonials. A testimonial is any comment from a past or current patient about the quality or result of your care. The ban covers direct testimonials, like a named patient review. It also covers indirect ones, like you retelling what patients say.

"Our patients consistently tell us" is you retelling what patients say. So it is an indirect testimonial. It also makes a claim about results ("significantly better"). That breaks a second rule on its own.

The line sounds like a friendly reassurance. If you did not know the rules, it would not worry you. That is exactly why it is risky.

The compliant alternative: Describe how you work, and leave patient feedback out. "Our assessment process is unhurried. We explain our findings and involve you in your treatment planning." This states a fact about your service. It does not hint at how patients felt or what result they got.

Violation 2: The outcome claim

This one showed up in the sports injury section:

"Our targeted rehabilitation programmes help athletes return to full training faster, with fewer setbacks."

AHPRA's rules also ban ads that create false hopes. In the rules' own words, an ad must not do anything that "creates an unreasonable expectation of beneficial treatment." That rule applies to all your ad copy, as well as patient quotes.

"Return to full training faster, with fewer setbacks" is a claim about results. It says your results beat someone else's, and it gives no proof. The obvious question is "faster than what?" The line suggests that if patients choose this clinic, they will get a better health result. The rules do not allow that.

Health marketing teams miss this one most often. Result-style wording is everywhere in sports physio marketing, because it feels natural. Athletes want to get back to training. Of course you want to say you help them do that.

But the way you say it matters. You can describe what your service involves. You cannot claim a certain result will follow.

The compliant alternative: "Our sports rehabilitation programmes are designed around the specific demands of your sport and training load. We work with your treating team to develop a graduated return-to-activity plan." This describes your service and your process. It does not promise a result or compare you with anyone.

Violation 3: The qualification statement

The third problem was subtle. Most small practices would miss it in their own review. It sat in the team section:

"Our lead physiotherapist is a specialist in sports and musculoskeletal physiotherapy with over 10 years of experience treating elite athletes."

In Australia, the word "specialist" is a protected title under AHPRA. A physio can only use it if they hold formal specialist registration with AHPRA. To get it, they must finish an approved specialist training program and pass an assessment. You cannot use it just because you have deep skill or long experience.

If you use "specialist" without that registration, the rules treat it as misleading. It does not matter how experienced you are. A physio with 20 years in sport and no specialist registration still cannot call themselves a specialist in any ad.

ChatGPT had no way to know the physio's registration. It used the word "specialist" because it is common in health marketing. It was copying patterns from general health content. It was not applying the rules.

The compliant alternative: "Our lead physiotherapist has worked in sports and musculoskeletal physiotherapy with both recreational and competitive athletes for over 10 years." It says the same thing without claiming a protected title.

Why does AI fail at compliance?

The mistakes above are not random. You can predict them from how AI tools are trained and what they aim for.

AI aims for smooth text, and accuracy comes second. An AI tool is trained to write text that reads naturally and matches what it has seen before. Rule-breaking lines often sound natural. That is why people write them. The human copywriters who wrote them first did not see a problem either.

AI doesn't know what it doesn't know. ChatGPT nodded to the AHPRA request. It wrote copy that sounded careful. But nodding to a request is very different from applying a complex set of rules. The tool had no reliable way to tell a safe claim from an unsafe one.

The rules need judgement in context. Is "patients feel better" a testimonial? That depends on the wording, the context and how a member of the public might fairly read it. Is "specialist" a protected title? That depends on how registration works in that profession. An AI cannot learn this from general internet text. It needs specific knowledge of the rules, applied to your page.

The training data problem. The internet is full of health content that breaks the AHPRA rules. Thousands of Australian health websites use testimonials, result claims and wrong title wording right now. The AI learned from that content. It learned what health copy looks like, mistakes and all.

So AI tools can still help with health content. They just cannot do the job alone. Someone who knows the rules needs to check what they write.

What should you do instead?

You can keep using AI. Just never let it write health content without a person checking it.

Here is a simple way to work:

Use AI for layout, and do the rule check yourself. AI tools are great at first drafts, laying out content and writing versions to test. Use them for that. Then check the copy against the rules before anything goes live.

Check for the 3 common mistakes. The 3 that cause the most trouble are testimonials (direct and indirect), result claims and title claims. For each piece of content, ask about each one by name. Skip the vague question "does this seem compliant?" Ask the clear ones: "does this contain a testimonial? Does this claim an outcome? Does this use a regulated title?"

Read the AHPRA guidelines yourself. AHPRA publishes its ad rules on its website and updates them from time to time. The document is specific and full of examples. If you write health content often, read it once a year. It is not long.

Give AI tools clearer instructions. Instead of "write AHPRA-compliant content," give it clear limits: "Do not include patient testimonials, direct or paraphrased. Do not claim any health outcomes. Do not use the word specialist unless the practitioner holds AHPRA specialist registration." Clearer limits give you better first drafts.

Work with a health copywriter for your most important pages. Your homepage, service pages and paid ads get seen the most. They are also the pages AHPRA is most likely to look at if someone complains. Getting them right from the start costs less than fixing them after a formal complaint.

If you write health content and want copy that wins patients without adding rule risk, take a look at how I approach this. Health ad rules are one of the areas I focus on. I work across physio, allied health, telehealth and specialist medical practices around Australia. Or get in touch directly to talk about your content.