A new survey shows Australian attitudes about the integration of artificial intelligence (AI) in government services.

Australians appear largely sceptical about the uptake of AI, and have registered a particular insistence on maintaining human oversight. 

The sentiment emerged from a study involving nearly 5,000 participants across Australia and New Zealand, conducted by Boston Consulting Group and Salesforce.

It found that only 40 per cent of respondents believe that the advantages of AI surpass its potential risks. 

Interestingly, Canberra residents displayed the most optimism towards AI, with a net positive view of 20 per cent, while those in the Northern Territory expressed the greatest concern, with 19 per cent having a negative stance.

Despite an increase in net satisfaction with digital government services in Australia by 7 percentage points, the survey unveiled that nearly two-thirds of users encounter basic usability issues, highlighting the complexity and navigational challenges across multiple agencies and government tiers for tasks such as registering births, marriages, building approvals, and more.

The federal government's approach to consolidating service and information around single portals was noted, contrasting with the quiet closure of the australia.gov.au site, which now redirects to the Prime Minister's website.

Another notable finding from the survey is the acceptance of AI by three-quarters of the respondents, contingent upon human supervision. 

This acceptance extends to AI's application in enhancing accessibility, language translation, and back-office processes, whereas its use in decision-making and sentiment monitoring raises concerns.

Amidst these findings, federal and state digital ministers recently agreed to a nationally consistent assurance approach, inspired by the NSW AI assurance framework, advocating for a new AI safety standard that mandates human involvement in government decisions.