A uni professor admitted using AI to write an opinion piece. Here’s what it revealed about trust in the technology

When a university vice-chancellor this week admitted to using AI in writing an opinion piece for a major Australian masthead, but did not disclose that use prior to publication, it highlighted the growing gap between people’s use of AI and trust in the technology.

Data from Roy Morgan this week showed 13.6m or 58% of the population older than 14 now use AI each month, with ChatGPT being the most popular, followed by Google’s Gemini and Microsoft Copilot.

Australians between the ages of 25-34 were most likely to use AI (74%) followed by 35-49 (72%), showing most of the workforce are those who are now using these tools.

Guardian Australia has covered extensively how it is affecting industries across the board – from healthcare to the legal system. As large language models are embedded into products we already use – whether we want it or not – people are going to use them.

Australians, however, deeply distrust AI, and how people use it. A survey from the Office of the Australian Information Commissioner last month found just 4% of Australians trust AI, putting the sector on par with data brokers, and 1% above social media platforms.

The survey found 79% of people want to know when AI is being used, up from 73% in 2023.

It’s the lack of transparency over when – and how – it is being used that seems to concern people the most.

The revelation this week that Western Sydney University pro vice-chancellor, Cath Ellis, had used AI in writing an opinion piece for the Sydney Morning Herald was less that she used AI, but more that the university admitted to it.

Western Sydney University admitted to Guardian Australia that Microsoft’s Copilot had been used in the development of the piece, which, ironically, made the argument arguing against students cutting corners as large language models are available to do essays and other university work.

The admission was made after the story went up, prompting the SMH to take down the piece after inquiries from Guardian Australia. The newspaper ran a mea culpa story later that day.

Had Ellis been transparent in the piece about her own use of AI and how it was useful, it might have been a different story. But then again, it may not have been published at all.

The availability of these models means this question of whether a piece of writing is truly the author’s own is unlikely to go away anytime soon, in academia, workplaces and media.

Fair Work Australia this week said it would ask for new powers to reject applications made with AI due to an unsustainable rise in applications that were clearly done with little effort or prospect of success.

Academic journals have instituted policies against generative AI being used in papers, but there are reports that AI use is growing rapidly.

There are even signs that in the world of programming – where AI is said to have brought the most benefit and efficiencies – there is pushback on its use in open-source programming like Zig.

Growing distrust in AI means now everyone can face witch-hunt-style accusations of using it, even if they may not be, while using AI can make people simply want to switch off.

Why would we bother to engage with something that people didn’t put much effort into?

Accusations that something was made with AI – but not disclosed – are becoming more and more frequent, and it isn’t just limited to news media.

The director of RuPaul’s latest film, Stop! That! Train! had to issue a statement this week declaring the film was not made using AI, after audiences who had seen early screenings believed some scenes were made with the technology.

Without transparency over how AI is being used, the benefits won’t be apparent, and it won’t be normalised as AI advocates may wish. Omitting its use out of the fear of being – correctly or incorrectly – called out for using AI, will just undermine trust in our existing industries.

The Guardian wp:paragraph