Way beyond robodebt: new $70m research centre to improve automated decision-making

By Stephen Easton

October 14, 2019

Getty Images

Researchers from eight Australian universities and 22 partner organisations from around the world will work to improve automated decision-making through a new centre of excellence based in Melbourne, starting out with about $70 million in funding.

The new Australian Research Council Centre of Excellence for Automated Decision-Making and Society is based at RMIT University and led by Professor Julian Thomas, who has recently been researching the implications of automated media and communication systems for consumers, citizens and governments.

“From artificial intelligence to the blockchain and big data, automated systems are changing our everyday life,” Thomas said.

“New systems offer enormous benefits in many areas but they also pose substantial risks to our privacy and security, and to our welfare as citizens and consumers. We urgently need a much deeper understanding of the potential risks of the new technologies, and the best strategies for mitigating these risks.

“Working with international partners and industry, the research will help Australians gain the full benefits of these new technologies, from better mobility, to improving our responses to humanitarian emergencies.”

Two federal ministers celebrated the Commonwealth contribution of $31.8m for the new centre last week, while elsewhere their colleagues and public servants continued to hotly defend Centrelink’s crudely automated robodebt system against compelling criticism in a Senate inquiry.

Minister for Education Dan Tehan said the new centre would bring together experts to “investigate responsible, ethical, and inclusive automated decision making” in a joint statement with Minister for Industry, Science and Technology Karen Andrews.

The kind of automated systems Tehan has in mind are way beyond robodebt in terms of technological sophistication — “self-driving cars or algorithms that are used to make medical diagnosis and business decisions” — but the welfare agency’s experience nonetheless demonstrates there can easily be a significant cost in terms of public opprobrium along with the expected financial benefit of automating a previously manual process. In the case of robodebt, the expected benefits have proven illusive.

Tehan said automation could “transform the efficiency” of industrial processes, as well as private and public service delivery.

“Our Government is funding research into automated decision making to ensure this technology provides the best possible outcomes for society and industry. Researchers will formulate world-leading policy and practice, inform public debate, and train a new generation of researchers and practitioners.”

RMIT University researchers collaborate with experts at seven Australian universities plus 22 academic and industry partner organisations from Australia, Europe, Asia and America, which are providing an additional $39.3 million in funding and “in-kind support” on top of the Commonwealth’s contribution.

Professor Jean Burgess, who leads the Queensland University of Technology Digital Media Research Centre, will be a chief investigator and the associate director of the new ARC centre of excellence, while several of her QUT colleagues will also be closely involved in the research.

“From AI to big data, automated systems are changing the way Australians live and work,” said Burgess. “They promise enormous benefits in many areas. But they also pose substantial risks to our welfare. We only need to look at the way automation is playing a major role in deciding what news we see and how online behaviour is managed by social media platforms.

“The new centre’s research program will provide a much-needed understanding of the risks of the new technologies, and the best strategies for mitigating these risks.

“We will work with a wide range of international research and industry partners, to help Australians gain the full benefits of these new technologies, from better health and transport, to improving algorithmic systems in news and media.

“The research will formulate world-leading policy and practice, inform public debate, and train researchers and practitioners in this challenging new cross-disciplinary field.”

Swinburne University researchers are also playing key roles. Professor Timos Sellis, the director of Swinburne University’s Data Science Research Institute, will co-lead a research program focused on effective and efficient use of data in automated decision-making and continue a prior focus on how to build fairness, accountability and transparency into automated systems.

The Dean of Swinburne Law School, Professor Dan Hunter, will co-lead a program focused on ethics in machine learning, building on his research in how AI can support fairer and more consistent criminal sentencing decisions.

About the author
Inline Feedbacks
View all comments
The Mandarin Premium

Canberra’s changed

Stay on top for only $5 a week


Get Premium Today