Curtin MP Kate Chaney says she is not convinced Australians will be protected by the Federal Government's recent promise to make the AI industry safer for the public.
Ms Chaney said that without specifics and legislation, the government's recent announcement that it would get tough on the AI industry rings hollow. "People here in Curtin are concerned about AI and the online deepfakes, emotionally intimate chatbots, supercharged scams, privacy concerns, workforce uncertainty and the unregulated automation of government decision-making," she said. "I welcome the minister highlighting a few of the 18 policy recommendations that I put forward in my AI discussion paper. "The devil will be in the detail."
Ms Chaney said it had been three years since the Robodebt Royal Commission recommended a legislated framework around automated decision-making. "This government talks a big game about never repeating Robodebt, but it hasn't put any policy in place to make this less likely," she said. In the absence of a government policy, she had developed a framework herself earlier this year to regulate automated decision-making in government. "This framework must be mandatory, and it must prioritise transparency and safeguards, such as human oversight and accountability for high-risk decisions," she said.
Federal Minister for Industry and Science Tim Ayres was one of six ministers who recently announced that the government would ensure the AI industry took safety measures to better protect the public. The ministers said they would legislate for AI companies to have a digital duty of care for consumers, beef up privacy protections for the public, and regulate how federal agencies used AI for automated decision-making. The laws would also address retail surveillance pricing the practice of retailers using AI to tailor online prices to what someone was willing to pay. They did not say when the new measures would be in place.
"Privacy reform should introduce a 'fair and reasonable' test for the use of personal data in Australia," Ms Chaney said. "This would mean that companies could only use personal data for fair and reasonable purposes, even if they've secured your consent hidden in the terms and conditions."