Almost three in four Australians want laws to restrict the use of generative AI tools and most people believe the technology is being developed too quickly, a survey has found.
But the university research, released on Monday, found less than half of all Australians had used the technology and most would like more media literacy training to become comfortable with it.
The findings come less than a week after a Senate inquiry into adopting AI heard calls for strict rules on the use of the technology amid warnings it could be used to spread misinformation online.
The Adult Media Literacy study, produced by Western Sydney University, the University of Canberra and QUT, surveyed more than 4400 Australian adults between January and April, asking questions about the media they used and their concerns.
It found 74 per cent of Australians wanted regulations to manage AI risks, 54 per cent thought it was being developed too quickly and 40 per cent worried generative AI would "harm Australian society".
But only 41 per cent of those surveyed revealed they had used generative AI tools, with 13 per cent saying they used text-based AI regularly and three per cent commonly using AI image tools.
Most Australians (80 per cent) said they wanted action to reduce misinformation online – six per cent more than in 2021 – and more than half of respondents said they wanted to learn more about how to identify false online claims.
Western Sydney University associate professor Tanya Notley said the research identified several groups which lacked confidence in their use of media, including those with lower levels of eduction, lower incomes and people living with a disability or in a regional area.
"Too many Australians have not received any form of media literacy education or they don't have access to support when they need it," she said.
"Most adult Australians are not confident about their ability to identify false and misleading information online, create a video and post it online, change social media privacy settings, or seek help from relevant authorities if they're being harassed online."
One in three Australians lacked confidence in their media abilities, the survey found, even though social networks were the most popular form of media in Australia, followed by television, streaming services, radio and news websites.
Independent senator David Pocock, who launched the report in Canberra, said the government should heed the public's concerns about misinformation and AI misuse.
Mr Pocock said many submissions to the Senate AI inquiry called for restrictions on risky AI use, which would be needed before the next federal election.
"Generative AI makes the 2016 'Mediscare' campaign look like absolute child's play," he said.
"The risk lies in the realms of deepfake democracy where a general population with low digital literacy, lacking confidence in their own ability to spot AI content, can be fooled."
Government officials on Friday told the Senate inquiry voluntary AI guidelines would be issued soon.
