Lead

The Sydney Morning Herald has retracted an opinion piece by a senior Australian academic after it emerged that the author used artificial intelligence to help write the article without disclosing that fact to readers. The piece, written by Professor Cath Ellis, pro vice-chancellor of quality and integrity at Western Sydney University, was published on the newspaper's website as a rebuttal to an earlier column by Macquarie University academic Kylie Moore-Gilbert.

Coverage Comparison

Reports from ABC Australia and The Guardian (both UK and Australia editions) agree on the core facts: Ellis used AI in drafting the piece, the Herald removed it after inquiries, and Western Sydney University confirmed the use of AI in a statement. The Guardian's coverage also included broader data on AI usage and trust in Australia, while ABC focused more narrowly on the incident and Nine's editorial guidelines.

Key Claims

  • AI use in writing: Multiple sources confirm that Cath Ellis used AI to write the opinion piece. According to a statement from Western Sydney University, Ellis uploaded 40,000 words of her own original materials into a Copilot Large Language Model, which summarised her knowledge and provided prompts for early drafts. The university described this as a "sophisticated and appropriate" use of generative AI.
  • Retraction due to non-disclosure: The Sydney Morning Herald took down the article after it was revealed that Ellis had not disclosed her AI use. Nine, the publisher of the Herald, has editorial guidelines that allow AI for research and ideas but not for writing stories, as reported by ABC Australia.
  • Ellis's argument: In the now-retracted piece, Ellis argued that universities were adapting to AI and that students should not "cut corners" or "outsource their thinking," according to The Guardian's report. The article was a response to Moore-Gilbert's claim that universities were committing "widespread, industrial-scale fraud" by allowing students to use AI.
  • AI detection: The Guardian reported that the column came up as 100% AI-generated when submitted to the AI-detector service Pangram. The university responded that such tools can detect AI use but cannot determine whether that use was appropriate.
  • Broader AI trust data: The Guardian's analysis cited data from Roy Morgan showing that 58% of Australians over 14 use AI monthly, while a survey from the Office of the Australian Information Commissioner found that only 4% trust AI and 79% want to know when it is being used. These figures were reported by a single source and have not been independently verified.

Perspectives

Academic integrity perspective: The incident raises questions about transparency in academic publishing. Ellis's use of AI, even to summarise her own work, was not disclosed to readers, which some may view as a breach of trust, especially given her position on academic integrity.

University's defense: Western Sydney University maintains that Ellis's use of AI was appropriate, arguing that it drew on her own expertise and experience. The university's statement emphasised that the technology was used to summarise her original materials, not to generate new ideas.

Media policy perspective: Nine's guidelines permit AI for research and ideas but not for writing stories. The retraction suggests that the Herald considered the non-disclosure a violation of these standards, highlighting the challenges media outlets face in regulating AI use.

Public trust perspective: The Guardian's broader analysis points to a significant gap between AI adoption and public trust. With most Australians using AI but few trusting it, the lack of transparency around its use—as in this case—may exacerbate concerns.

Conclusion

The retraction of Cath Ellis's opinion piece underscores the growing tension between the use of AI in professional and academic settings and the public's demand for transparency. While the university defends the use as appropriate, the lack of disclosure led to the article's removal. As AI becomes more embedded in daily life, the incident serves as a reminder of the importance of clear guidelines and honest communication about its role in content creation.