This week, a university’s professional vice-chancellor admitted to using AI to write an opinion piece for Australia’s main mastheads, but did not disclose that use before publishing, highlighting the widening gap between people’s use of AI and their trust in technology.
According to data released this week by Roy Morgan, 13.6 million people, or 58% of the population over the age of 14, are currently using AI every month, with ChatGPT being the most popular, followed by Google’s Gemini and Microsoft Copilot.
Australians aged 25-34 are most likely to use AI (74%), followed by those aged 35-49 (72%), indicating that the majority of employees are currently using AI tools.
Guardian Australia has extensive coverage of how the virus is impacting industries across the board, from healthcare to the legal system. Large language models are built into the products we already use, so people are going to use them whether they want to or not.
However, Australians have a deep mistrust of AI and how people use it. A survey conducted by the Australian Information Commissioner’s Office last month found that only 4% of Australians trust AI, putting the sector on par with data brokers and 1% more than social media platforms.
Research shows that 79% of people want to know when AI will be used, up from 73% in 2023.
What people are most concerned about is the lack of transparency regarding how and when it is being used.
This week’s revelation that Western Sydney University’s professional vice-chancellor, Cass Ellis, used AI when writing an opinion piece for the Sydney Morning Herald was not so much that she used AI, but rather that the university acknowledged it.
Western Sydney University confirmed to Guardian Australia that Microsoft’s Copilot was used to develop the work, but ironically, the large language models available to do essays and other university assignments have led to arguments against students cutting corners.
SMH acknowledged this fact after the article became popular, and following inquiries from Guardian Australia, SMH was forced to remove the article. Later that day, the newspaper published a serious crime story.
If Ellis had been transparent in her article about her own use of AI and how it helps, things might have been different. However, it may not be published at all.
The availability of these models means that this question of whether the text is truly the author’s own is unlikely to disappear in academia, the workplace, and the media any time soon.
Fair Work Australia announced this week that it would seek new powers to refuse applications made using AI due to an unsustainable increase in applications made with apparently little effort or chance of success.
Although academic journals have policies prohibiting the use of generative AI in articles, there are reports that the use of AI is rapidly increasing.
Even in the programming world where AI is said to have brought the greatest benefits and efficiencies, there are even signs of a backlash against its use in open source programming like Zig.
Growing distrust of AI means that anyone who uses it could face accusations of witch-hunting even if they don’t use it, but it also means that using it could make them want to simply switch off.
Why bother working on something that people don’t put much effort into?
Accusations that something is created with AI but not made public are becoming more frequent, and they are not limited to news media.
RuPaul’s latest film, Stop! that! train! The film was forced to issue a statement this week saying it was not made using AI after early screenings led audiences to believe that some scenes were made using AI.
Without transparency about how AI is used, its benefits will not be clear and it will not be normalized as AI advocates hope. Skipping the use of AI for fear of being criticized for right or wrong will only erode trust in the existing industry.
