30 SEPTEMBER, 2026
Who: AI expert Professor Niusha Shafiabady, who is an adjunct with Charles Darwin University Adjunct and Head of Discipline for IT and the Director of Women in AI for Social Good lab at Australian Catholic University.
Topics:
-
AI detection software and its use in creative publishing, following multiple high-profile instances of AI accusations in the industry.
Contact details: Call +61 8 8946 6721 or email [email protected] to arrange an interview.
Quotes attributable to Professor Niusha Shafiabady:
"AI detection software should never be treated as definitive proof that an author has used artificial intelligence. These systems analyse patterns in language and generate a classification based on those patterns. They do not observe how a manuscript was actually created. A detection result should therefore be the beginning of further inquiry, not the conclusion of it.
"There is also a genuine risk of bias. Research has shown that some AI detection systems misclassify human writing, including writing by people for whom English is not a first language. This matters because writing style is shaped by many influences, including language background, culture, education, genre and individual expression. Some of these characteristics may overlap with patterns that a detection system associates with machine‑generated text.
"At the same time, it is important to be cautious about attributing racial bias to any specific case without clear evidence. Bias can enter the process at multiple points. It may arise from the data used to train a system, but it can also arise from human decisions about whose work is examined, who is suspected, how a detection result is interpreted and whether standards are applied consistently across authors.
“This is why the entire human and AI decision‑making chain needs to be considered, rather than focusing solely on the software.
"Publishers should not rely on an AI detection score in isolation. They should consider other forms of evidence, such as earlier drafts, document histories, metadata, editorial correspondence and the development of the manuscript over time. Authors should also be given an opportunity to explain their writing process before any conclusions are reached.
"My view is that AI detection can contribute information to an investigation, but it should never replace human judgement, transparency and due process, particularly when a person’s reputation and career are at stake."
Contact details:
Raphaella Saroukos she/her
Research Communications Officer
Marketing, Media & Communications
Larrakia Country
T: +61 8 8946 6721
E: [email protected]
W: cdu.edu.au
Charles Darwin University acknowledges all First Nations people across the lands on which we live and work, and we pay our respects to Elders both past and present.
CRICOS Provider No. 00300K (NT/VIC) I 03286A (NSW) | RTO Provider No. 0373 | TEQSA Provider ID PRV12069
