A senior academic at Western Sydney University employed artificial intelligence to compose an opinion piece that encouraged students to avoid relying on AI for their studies, sparking controversy and resulting in the article’s removal from the Sydney Morning Herald website.

Professor Cath Ellis, pro vice-chancellor for quality and integrity at Western Sydney University, authored the opinion piece in response to comments by academic Kylie Moore-Gilbert. Moore-Gilbert had expressed concern over students potentially outsourcing their learning to AI, advising her stepdaughter to reconsider university enrollment amid a changing educational landscape where students seemed to be assessed on their ability to craft effective AI prompts.

Ellis acknowledged the challenges posed by AI but maintained that students should engage fully with their education. She urged learners to “do the work” and resist shortcuts, arguing that genuine effort would be distinguishable even in a system disrupted by emerging technologies.

However, the article was flagged by the AI-detection service Pangram as likely generated by artificial intelligence. Following inquiries, Western Sydney University confirmed that Ellis had used a large language model (LLM), specifically Copilot, to produce her piece. According to a university spokesperson, Ellis provided the AI tool with 40,000 words of her own original material, which the model summarized to generate prompts that formed the basis of initial drafts. The university defended this approach, describing it as a “sophisticated and appropriate use” of generative AI that reflected Ellis’s extensive expertise and longstanding scholarship in the field.

The spokesperson emphasized that AI detection tools cannot assess whether AI use is appropriate or otherwise and stated that the university considered the application of AI in this instance to be suitable.

The Sydney Morning Herald’s parent company, Nine, has editorial policies permitting writers to use AI for preliminary research and ideation but explicitly prohibits AI from being used to generate stories for publication. When AI-generated content is published, it must be clearly labeled, although the use of assistive AI does not require declaration under the policy. Nine did not respond to requests for comment.

Sydney Morning Herald editor Jordan Baker later confirmed the article was removed after the publication discovered that neither Ellis nor Western Sydney University disclosed the use of AI in writing the opinion piece. Baker described the omission as “unacceptable” and stated that an investigation was underway.

This incident adds to growing concerns about transparency in the use of AI in journalism and academia, as the integration of generative AI tools becomes more widespread globally. Similar controversies have emerged recently in media circles, including a freelance journalist’s admission of using AI to produce a book review that bore similarities to previously published content, leading to the New York Times severing ties with him.