Australian employer, business tech and research groups have formed an alliance to tackle responsible AI by developing best practice guidance, tools and training for Australian companies using the high-risk technology.
The Responsible AI Network to be launched on Thursday includes the industry groups alongside standards, ethics and research organisations. No consumer or worker groups are yet involved, with the focus initially on businesses, but more partners are being sought.
It will offer Australian businesses support with guidance and training on the fundamental challenges of AI like data bias and unknown or unforeseen outcomes so business users build explainability, fairness and accountability into Australian AI systems.
“The fact that we can never really know the full outcome and impact is what makes responsible AI such an elusive domain,” Australia’s National AI Centre director Stela Solar said.
With a mix of laws, regulations and guidance surrounding AI, and more standards rapidly approaching, the new group will work to prepare Australian business.
“There’s a wave of standards coming from the international standards organisation, which we believe is going to catch the commercial sector by surprise, and that’s who we really want to support,” she told InnovationAus.com.
“AI technologies are already transforming industries, but they must be adopted in
Industry minister Ed Husic will officially launch the Responsible AI Network on Thursday. He said the increasingly popular technology must be adopted in “ways that work for people, not against people”.
“AI shouldn’t be seen as just a business tool. We need to put it to work to benefit communities and national wellbeing too,” Mr Husic said.
The new network will look to make Australia a global leader in responsible AI, said Ms Solar, a former Microsoft global head of AI solution sales who returned to Sydney in 2021 to launch CSIRO’s AI push.
“The core reason why [responsible AI] is so elusive is there are a lot of variables in an AI system all the way from the data it’s built on to the outcomes it creates. And we’re only just at that tip of the iceberg or really understanding the full power of scale of the AI technology.”
The Responsible AI Network will be launched by Industry and Science minister Ed Husic on Thursday ahead of an event with Standards Australia next month and the opening of an online hub for Australian businesses to access the advice.
Initial partners of the Responsible AI Network include the Australian Industry Group, Australian Information Industry Association (AIIA), CEDA, CSIRO’s Data61, Standards Australia, The Ethics Centre, The Gradient Institute, The Human Technology Institute, and the Tech Council of Australia.
The network is centred around six core pillars: law, standards, principles, governance, leadership and technology.
Businesses are being invited to join as members to form a community of practice and access the network’s knowledge, guidance, best practice and frameworks to get them to a ”base level” of responsible AI.
Ms Solar said while guidance will be the initial focus, the network may eventually develop more structured approaches like a voluntary code on responsible AI.
With no nation yet to “work out” responsible AI, Ms Solar said, a massive opportunity exists for Australia, which already has a strong AI research base.
She said Australia has all the required elements to lead the world on responsible AI, including AI Ethics principles, early investment in responsible AI research, the world’s first online safety regulator, and a “fair go” being “core” to Australian values.
“We have the ingredients to now translate this capability into industry and really step into this opportunity and help our industry uplift.”
Mr Husic agreed the talent exists but needs to be scaled up “for the longer benefit of the nation”.
“We want Australia to be a leader in this area, but, we also need to be taking up these new technologies in safe and inclusive ways. That’s really important,” he said.
