AI Is Writing Your Competitor's Content. Here's What Google Actually Does About It.

Google does not penalise AI content in healthcare just because AI wrote it. It says so in writing. Its guidance says it focuses on "the quality of content, rather than how content is produced", and that "Appropriate use of AI or automation is not against our guidelines". What Google does punish is scaled content abuse. That means lots of low-value pages made to game search, whoever typed them. AI detectors also make poor proof. In one study, 7 detectors wrongly flagged an average of 61.22% of essays by people writing English as a second language. For your clinic, the real risk sits somewhere else. It sits in section 133 of the National Law, the health ad rules.

Can Google tell if your competitor used AI?

Maybe. But you are asking the wrong question, and Google has said so for years.

Google's guidance on AI content is public, so you can read it yourself. This is the line that matters: "Our focus on the quality of content, rather than how content is produced, is a useful guide that has helped us deliver reliable, high quality results to users for years."

The FAQ on the same page puts it even more simply: "Appropriate use of AI or automation is not against our guidelines."

The main text says the same thing: "it's important to recognize that not all use of automation, including AI generation, is spam".

Google also tells you why. It points back to a time when lots of human writing was mass-produced too. In its words: "No one would have thought it reasonable for us to declare a ban on all human-generated content in response."

So if you are waiting for Google to punish the clinic down the road, you will wait a long time. Google has ruled that out in writing. It is worth knowing before you spend money on the wrong fix.

What does Google actually penalise?

Google punishes something it calls scaled content abuse. It is a named rule in Google's spam policies, and the definition is short.

"Scaled content abuse is when many pages are generated for the primary purpose of manipulating search rankings and not helping users."

The next sentence does the real work: "This abusive practice is typically focused on creating large amounts of unoriginal content that provides little to no value to users, no matter how it's created."

Read those last 5 words again: no matter how it's created. The rule does not care which tool you used. A person can break it with a keyboard, and so can an AI tool.

The policy then gives examples. The first one is the trap clinics fall into most: "Using generative AI tools or other similar tools to generate many pages without adding value for users".

On a clinic website, this nearly always looks like suburb pages. You might have Physiotherapy in Carlton, Physiotherapy in Fitzroy and Physiotherapy in Brunswick. Soon you have 30 pages where only the suburb name changes.

That pattern was a risk before anyone had heard of ChatGPT. AI just made those 30 pages cheaper to make. The rule itself has not changed.

Does Google treat health content differently?

Yes, and Google says so on the same page.

Google names health as 1 of 3 topics where good information really matters. The other two are civic and money topics. For these, it says, "our systems place an even greater emphasis on signals of reliability".

That one sentence explains why your clinic cannot treat content the way a homewares shop can. By Google's own account, the bar is higher for your topic.

It also matches what gets a clinic page quoted in AI answers. We covered that in why ranking on Google is not enough. Trust signals are things like your practitioners' names, registration numbers, real qualifications, dates and sources. An AI tool does not know these about your practice, and it will get them wrong if it guesses.

Can anyone prove a competitor used AI?

Not reliably. Trying can also backfire in a way that matters for Australian clinics.

Researchers tested 7 popular AI detectors on 91 essays written by real people sitting the TOEFL exam. These were people writing English as a second language. They also tested 88 essays by US 8th-grade students.

On the US 8th-grade essays, the detectors were "near-perfect", in the study's words. On the TOEFL essays, they fell apart.

So the tool you might use to prove a competitor cheated often blames human writers instead. It also picks on one group. It misfires on people who write English as a second language.

That matters more in health than in most fields. The Ahpra annual report for 2024/25 says 26,703 practitioners trained overseas gained registration to work in Australia that year alone.

Training overseas does not mean someone writes English as a second language. Plenty trained in the UK, Ireland, New Zealand and Canada, and the report does not split the group by language. So treat this as an overlap, and do not assume the two are the same. The overlap is still big enough that running a detector over your team bios is a bad idea. The most likely result is that you accuse a real practitioner of not writing their own page.

Who is liable when AI writes a clinic's advertising?

This part has nothing to do with Google. It is the part that carries a real legal penalty for you.

Section 133 of the National Law (the law that covers registered health practitioners) is short. It says you must not advertise a health service in a way that does any of these 5 things:

• Is false, misleading or deceptive.
• Offers a gift, discount or other deal without stating its terms.
• "Uses testimonials or purported testimonials about the service or business".
• "Creates an unreasonable expectation of beneficial treatment".
• Pushes people to use health services they do not need.

The maximum penalty is $60,000 for a person and $120,000 for a company.

Now read subsection (2). Almost nobody quotes it:

That is a defence, and look who it protects. It covers whoever makes or publishes the ad for someone else. That could be the printer, the platform, or arguably the agency.

Your practice, the one being advertised, gets no matching defence. Years before AI tools existed, the law decided who carries the risk. It is you.

So you cannot say "the AI wrote it" as a defence. You cannot say "the agency wrote it" either. We went through that in detail in who is liable when an agency writes your copy.

This really happens. We asked ChatGPT to write compliant copy for a physiotherapy clinic. It broke the rules in 3 separate places, and the copy read beautifully. You can read the full test in what happened when we asked ChatGPT for AHPRA-compliant content.

The reason is simple. An AI tool writes what marketing copy usually looks like. Most of the copy it learnt from comes from countries where patient reviews, promised results and "best in the area" are all legal. In Australia, they are not.

What does a clinic page need to survive both tests?

Two judges read your page, and they want different things. Your page has to pass both.

Google wants a reason for the page to exist. On a health topic, it also wants signs you can be trusted. AHPRA, the body behind the health ad rules, does not care about your rankings at all. It cares whether your words break section 133.

Your page passes both when it holds things an AI tool cannot make up about you:

• A named practitioner, with their registration number and their real qualification.
• What the appointment involves, in detail, including what it does not cover.
• Real prices, or a real reason you do not show them.
• Dates, and a source for any clinical claim.
• Leave out patient quotes, promised results, and words like "leading" or "best".

Notice that every item on that list also builds trust with Google. That is the good news. The work that keeps you within the rules is mostly the same work that makes your page worth ranking. In health, following the rules and writing well pull the same way. That is not true in every industry.

So what should you actually do?

Stop watching the competitor. Their AI content is not your problem, and you cannot prove it anyway.

Do these 4 things instead.

One. Count your own thin pages. Look for suburb or service pages where only one word changes. That is the scaled content abuse pattern, and the risk is on your site, not theirs. Merge them into one page that is really good.

Two. Read your own site against section 133 before a regulator does. AHPRA looked at 775 advertising complaints in 2024/25. It has also trialled AI tools that scan websites. We covered that in what 775 advertising complaints actually led to. AHPRA no longer waits for someone to complain.

Three. Keep using AI, but change the job you give it. Use it for structure, first drafts, and turning a practitioner's spoken answer into clean writing. Do not let it supply facts about your practice or any clinical claim. That split passes both tests.

Four. Put the human evidence back in. Add names, numbers, dates and sources. Google says it weighs these more heavily on health topics. They are also things an AI tool cannot fake about you.