On May 27, Kylie Moore-Gilbert, a political scientist and writer at Macquarie University in Sydney, Australia, published an article in the Sydney Morning Herald titled “I’m an Academic, but I’ve Told My Stepdaughter to Think Twice About Going to University.“
In the article, Moore-Gilbert says that AI may be harming the value of a degree as a credential because future employers will be looking for better prompt writers, rather than an educational background.
A week later, Cath Ellis, the pro vice chancellor of quality and integrity of Western Sydney University, published a rebuttal in the same paper. In it, she encouraged students to “do the work” and not cut corners.
However, the column became famous for less-than-desirable reasons. Some readers began to suspect that the article was written using AI, noting AI phrasing throughout the piece. Others submitted the column to AI-detector services, including Pangram, which flagged it as AI-generated.
Ellis, for her part, didn’t deny using AI to assist in writing the article. She said she uploaded 40,000 of her own words on the subject into a Copilot Large Language Model (LLM), which then summarized her thoughts and knowledge.
The University also used AI to edit and proofread the article before publication.
The Sydney Morning Herald pulled the article from its website, citing its policies against AI usage. Ellis said that she was not aware of that policy and would not have used AI if she had known about it. However, her university released a statement saying that her use of AI was “sophisticated and appropriate.”
This is because Western Sydney University, as a school, has largely embraced the use of AI. The school encourages students to use AI when allowed under the guidelines.
But, while it might have been appropriate at the school, it was not appropriate for the environment it was published in. Though debates about the role of AI in education are ongoing, it’s not the school’s rules that were the most important.
Disclosure: I do know Cath Ellis professionally, we have met at several conferences and we are in the same network on LinkedIn. However, I’ve not worked any projects with her or Western Sydney University.
Back in February, I wrote a column entitled “If You Don’t Disclose AI, You’re a Plagiarist.” While that is still true, what disclosure is needed and when depends heavily on the context of what is being created, the audience it is being distributed to, and the norms of that particular environment.
This is why I keep my own AI disclosure, even for the things that I cannot control.
The main tension here is simple. Western Sydney University has a relatively permissive AI policy. The Sydney Morning Herald does not. When Ellis wrote an article for the Herald, she should have reviewed and understood those standards.
Ellis said that she was unaware of the policy. If that’s the case, I will say that the paper should make greater efforts to ensure that its guest authors are made aware of it. However, both Ellis and the school had an obligation to ensure their submission complied with all of the paper’s policies, including on AI usage.
I’ve talked about this before in the context of citation. Basically, when you switch to a different environment, you need to ensure that your citations follow the norms of that space. You don’t try to force your existing standards and norms on the new environment.
Academics expect this of their students. If you come into a classroom and cite material the way you would in a social media post or a casual letter, you will be marked down and likely find yourself facing allegations of plagiarism.
However, some forget to extend the same grace when they step outside of academia. From a purely practical standpoint, that is the issue here. A practice that the university found acceptable was not acceptable at the paper. But the paper is the final arbiter of what is and is not acceptable on their pages and in their site.
That is why they took the article down.
It would be easy to leave this here. It would be much simpler to say that the school and the paper had different policies around AI usage and that Ellis violated the paper’s policy while following her school’s. However, she should have been using the paper’s policy for this particular project.
But there is another layer to this, one that is obvious to everyone who has been following it. Ellis’ paper wasn’t just any paper. It was a guest column encouraging students to not cut corners and to do the work. She was making a pitch that college degrees are still valuable in the age of AI.
I agree strongly with that message. And I think that the usage of AI in making that point could have been a powerful statement. You could show how AI can supplement and aid but not replace the knowledge and skills that a degree provides.
For an AI-permissive school like Western Sydney University, that could have been a powerful message. Many people would have still frowned upon it, including me, but it would have had a consistent message that adds to the ongoing debate around AI.
However, to make that point, the AI usage would have needed to be disclosed. Whether in the narrative of the article or in a disclosure statement, the reader would have needed to know in advance that the article was written using AI.
With no disclosure, that point is completely lost. Then, when the AI usage is uncovered later, it becomes a moment of extreme irony. Part of being involved with the conversation around AI usage means being transparent about your own usage.
People who are writing about this topic need to assume they are living in a glass house. This means being honest about when, how, and where they are using AI. Just because a usage is acceptable doesn’t mean that it shouldn’t be disclosed.
It is better to over-disclose than under-disclose. That has been true of traditional citation issues for centuries and is true of AI usage today.
Right now, the conversation around AI is a complete mess.
First, there’s no good definition for what is and is not AI. AI is more of a marketing term than a meaningful moniker. Just because a tool uses an LLM doesn’t mean that its function couldn’t have been achieved without one. Even I have no idea what is and is not “AI” in 2026.
Second, the vast majority of people using AI tools are not disclosing their usage. This, coupled with the lack of truly meaningful AI detection tools, means that there is no way to know what is or is not created using generative AI.
Finally, the objections over AI usage run a gamut. There are environmental concerns, ethical concerns over the use of human-generated content to train AI systems, fears over job displacement, and, of course, authorship issues. It’s a fragmented argument.
Western Sydney University has chosen to largely embrace the use of AI. That is their choice. Whether you agree with it or not, it is their choice. They get to decide the rules on their campus.
However, making that choice doesn’t mean that they can enforce their own standards in other spaces. Ellis and the school had an obligation to respect the wishes of the paper. If they didn’t know, they should have asked. It was a major blunder made worse by the topic.
If you are using AI, regardless of the capacity, please don’t get so comfortable with your usage that you forget others might object to it. Always disclose what you do and make sure that your audience is aware of it. What you see as innocent or inconsequential might be seen by someone else as a major infraction.
But, if you aren’t willing to disclose it, then you shouldn’t be doing it. That’s true for just about anything, not just AI.
If you want to feature this article in your site, classroom or elsewhere, just let us know! We usually grant permission within 24 hours.
Click Here to Get Permission for Free
I am not a lawyer. I am just a legally minded Webmaster/Writer frustrated by the plague of plagiarism online and doing something about it.
AI Disclosure
Need a Plagiarism or Copyright Expert for a news story? Contact Jonathan Bailey here.
Plagiarism Today is a website aimed at webmasters and other creators to help them address the issues of plagiarism, copyright infringement and other forms of content misuse online. I am not a lawyer and nothing on this website should be construed as legal advice.
Content Removal Policy
Leave a Reply