These measures are the government's interim response to a review carried out by the Department of Industry, Science and Resources, which led to the creation of a 12-member ad hoc expert panel on AI, including many academics.
One of the topics the agency is considering is a model for future AI laws, which could be applied to a wide range of industries deemed high-risk, including healthcare, finance and housing. They can ban specific practices, such as creating social scores for customers, as the European Union does, or they can apply general standards, such as non-discrimination.
People familiar with the government's plans say the new advisory body will be permanent and include more business representatives, as companies will be the primary adopters of AI tools. Business bodies, including the Business Council of Australia, are only observers of the current structure.
Industry and Science Minister Ed Husik declined to comment directly on the permanent expert group, but he did not comment on how the ad hoc advisory group would define “high-risk” artificial intelligence and what harm could come from it. He said he is working on ways to find out if there is.
“There are also options for mandatory guardrails for high-risk AI, which could include more testing and transparency, or whether some types of very high-risk AI might require an outright ban. We are also investigating,” Husich said. “We look forward to discussing the group's findings in more detail in the coming weeks and months.”
Proponents of artificial intelligence point to its ability to increase productivity by automating white-collar tasks and facilitating the creation of new services. But detractors see it as a potential vehicle for the spread of discrimination and political bias, especially when deployed in areas such as health care, finance, housing, and the media.
A survey of 46 technology leaders conducted by Datacom and the Australian Tech Council, which represents major technology companies and emerging startups, found that artificial intelligence will be the top industry trend this year.
Industry bodies such as the TCA and Business Council of Australia will welcome the government's plans to increase business influence over AI. The restructuring of the advisory committee is one of the Technical Council's requests for the federal budget, according to a filing seen by . Australian Financial Review This also requires governments to maintain a flexible approach to AI.
“It is important to recognize that we are not starting from scratch,” said TCA Acting Administrator Ryan Black. “Australia has a sound existing legal framework relating to AI and a specialized regulator that regulates products using AI, including in areas such as healthcare and financial services.”
Mr Black, who represents major technology companies and emerging startups, said building on existing standards and regulations would help Australia maintain a competitive economy as AI adoption soars.
The European Union is artificial intelligence method It would prohibit practices such as using AI to assign “social scores” based on people's behavior, as China has done.
The rules will be implemented in stages, in contrast to countries such as the UK, where Chancellor Rishi Sunak has declared the country to be “pro-innovation”. As part of this, applying existing anti-discrimination rules to AI avoids the problem of creating specific standards that can be rapidly rendered obsolete by technology.
Australian Council of Trade Unions ACTU secretary Sally McManus said workers should receive a fair share of the productivity gains brought about by AI.
“Worker voices must be at the center of adoption and there must be clear protections to ensure it does not erode living standards or roll back workers' rights,” McManus said in a statement. said.
“The use of AI has already been demonstrated to have risks leading to discrimination and unfair decision-making by employers.”
BCA chief executive Bran Black (no relation) praised the potential of AI.
“The use of artificial intelligence will make our businesses more competitive and productive, which is good for growth, investment and job growth in Australia,” Mr Black said. He supported the government's previously announced plans for a more principles-based approach.
Australia does not yet have a breakthrough artificial intelligence company, but the technology is widely applied by existing companies to help guide decision-making, analyze data and answer customer service questions.
Luke Latham, Australian general manager of fintech Airwallex, acknowledged his excitement about AI but cautioned that businesses should be realistic in their adoption.
“The effectiveness of AI ultimately depends on the underlying data that informs these tools,” Latham said in a statement.
“At the end of the day, the output is only as good as the input, so investing in getting the foundations right is essential, without which companies will struggle to produce such ambitious results.” I will.”
There have already been several cases where AI companies have violated existing laws, illustrating the potential risks and how current laws can respond.
In one case, facial recognition company Clearview AI was found to have violated the privacy of Australians by collecting their images.
Travel website trivago has been ordered to pay $44.7 million in 2022 for using algorithms to trick users into booking hotels at higher prices.
