Jordan Baker, the Herald’s editor, said the paper removed Prof Cath Ellis’s column after learning generative AI had been used without disclosure, calling that omission “unacceptable” and saying an investigation is under way.
100% AI-generated was the result when the article was run through Pangram, even though the piece had urged students not to “outsource your thinking” to AI.
Western Sydney University said Ellis fed 40,000 words of her own material into Microsoft Copilot, arguing the model only summarized her expertise and that the AI use was appropriate.
Nine’s editorial policy allows AI for research and idea prompts but says AI will not be used to write stories for publication; AI-generated material must be clearly labeled.
The removal adds to a broader run of undisclosed-AI publishing incidents, as newsrooms and universities face growing scrutiny over how generative tools are used.
Is using AI to summarize your own work a clever tool or an unacceptable breach of editorial trust?
If experts secretly use AI to write, how can readers ever trust the source of information again?
The Fallout from SMH’s AI-Generated Opinion Piece: How Academia and Journalism Are Responding to the Transparency Crisis
Overview
On June 2, 2026, The Sydney Morning Herald removed an opinion piece by Professor Cath Ellis after it was revealed that artificial intelligence played a major role in its creation. Professor Ellis began by uploading 40,000 words of her own work into a Copilot Large Language Model, which summarized her knowledge and generated prompts for the article’s early drafts. She then shared the draft with her university’s media team, who also used AI tools to suggest improvements. The article went through several revisions in this collaborative human-AI process, highlighting the growing impact of AI on content creation and the importance of transparency.