A top Sydney academic has written an opinion piece using AI to encourage students to use such technology to cut corners and “get to work”, and the Sydney Morning Herald has removed an “unacceptable” article from its website.
Western Sydney University’s professional vice-chancellor for quality and integrity, Professor Cath Ellis, published an opinion piece in the Sydney Morning Herald last month in response to an article by academic Kylie Moore-Gilbert.
Moore-Gilbert wrote that he advised his stepdaughter to think twice before enrolling in college because students can easily outsource their learning to AI, and said that students are “being judged on who can write the best AI prompts.”
In response, Ellis wrote in an article that while “the problems with AI are real,” students should still go to university and study properly.
“Don’t cut corners. Don’t outsource your thinking, no matter how tempting it may be. If the system is as weak as some say, real effort will not be hidden; it will be visible,” she wrote.
However, when the column was submitted to AI detection service Pangram, it appeared as generated by AI.
In response to questions from Guardian Australia, the university said Mr Ellis used AI to write the column.
“Professor Ellis uploaded 40,000 words of her own original material to the Copilot Large Language Model (LLM) to write the opinion piece. The model summarized her extensive knowledge base and provided prompts,” the spokesperson said.
“This is the basis of an earlier draft and reflects Professor Ellis’s own thinking, ideas and opinions, accumulated over more than a decade of dedicated research as a world leader in this field.”
The spokesperson said using the LLM to leverage its own expertise and experience “demonstrates a sophisticated and appropriate use” of generative AI.
“While programs like Pangram can detect the use of AI, they cannot determine whether that use was appropriate or inappropriate,” the spokesperson said. “The university believes that this use of AI was appropriate.”
As of Wednesday morning, Ellis’ opinion piece did not explicitly state that it used generative AI to write the article.
The Sydney Morning Herald’s parent company Nine’s editorial policy allows writers to use AI for initial research and ideation, but states that “AI will not be used to write articles for publication.”
The policy states that places where AI-generated material is published will be clearly labeled, but that “the use of assistive AI does not require a declaration.”
Nine did not respond to Guardian Australia’s questions. However, in a subsequent article published on Wednesday, Sydney Morning Herald editor Jordan Baker said the article did not meet editorial guidelines and was removed.
“The Herald was not informed by either the author or Western Sydney University about the use of AI in editing the article,” Mr Baker said.
“This is clearly unacceptable and we are investigating further.”
As generative AI tools become embedded in workplaces around the world, incidents of news organizations disclosing output from AI are becoming more common. In March, Crikey removed a series of articles by a writer after it was revealed that he had used AI to help proofread his copy.
Also in March, The New York Times cut ties with a freelance journalist after he admitted to using AI to write a book review that mimicked the content of a book published in The Guardian.
University commencement speakers who praised AI have been booed by students in recent weeks amid concerns that their job prospects will be dashed as companies replace graduate students’ jobs with AI.
