Source: The Guardian

Federal departments and agencies that use automated decision-making will face stringent new rules under a national AI plan that aims to extend protections into consumer rights, workplace safety and privacy. Senior ministers are developing requirements to build safety into government AI systems, emphasising fairness, accuracy and transparency, with the attorney general, Michelle Rowland, assigned to lead the rulemaking work. The government is also pursuing legislation to establish a digital duty of care to place obligations on AI companies to design in safety, an initiative led by communications minister Anika Wells. The prime minister has announced a new office of AI within his department as part of a whole-of-government approach, and the plan includes a world-first standard for large, energy-intensive datacentres expected to be legislated from next year to govern locations and community impacts.

Federal use of automated decision-making in service delivery remains significant and raises concerns that more sophisticated AI could supplant human-led processes, a vulnerability underscored by the previous robodebt failures. The assistant minister for science, technology and the digital economy, Andrew Charlton, warned about rapid datacentre growth and high electricity consumption, saying the proposed datacentre standard will aim to prevent local harms such as proximity to homes and schools, price impacts and water use. A second round of privacy law reforms is planned to strengthen and simplify personal data protections. Opposition figures, including deputy Liberal leader Jane Hume, criticised the government’s focus on datacentre rules as insufficient to address the broader set of AI risks.