According to Lianhe Zaobao citing ABC, the Australian government will tighten rules on federal agencies’ use of artificial intelligence to ensure decisions involving welfare and public services are fair and accurate, while strengthening protections for personal data and consumer rights. Federal agencies including Centrelink and Services Australia will have to follow a new framework when using AI to assist decisions, with proper oversight required.
The move comes after years of pressure to keep human involvement in areas affecting public benefits while using AI to improve administrative efficiency. Australia’s federal ombudsman last year found that income support payments for nearly 1,000 job seekers were unlawfully suspended because of an automated cancellation mechanism. The government’s use of algorithms in home-care assessments and budgets under the National Disability Insurance Scheme has also drawn criticism, and the ombudsman began investigating the aged-care assessment tool earlier this year.
Employment Minister Rishworth said in May, when announcing reforms to unemployment assistance, that humans would continue to be involved in assessing job seekers’ payments. The government also plans to review consumer and privacy protections, including so-called surveillance pricing, where retailers collect behavioral data and set different prices for different individuals. It will examine how to strengthen oversight through consumer law and whether AI firms should be brought under “digital duty of care” legislation requiring platforms to identify and address potential harm. Prime Minister Albanese has said workers must help shape the AI transition and that AI firms will not be allowed to train models on Australian books, music and art without compensation.