Why Australia is falling behind on AI

By Australia Financial Review

Hitesh Choudhary T1 Pa Ib Mtjim Unsplash

Experts have warned an immature legal framework, a lack of technologists on company boards, and poor collaboration between business and academia is preventing Australia from realising the full potential of rapid developments in artificial intelligence technology.

Speaking at The Australian Financial Review Future Briefings event in Sydney, a panel of experts from across related disciplines said that while a flurry of AI development is underway across local universities, law firms, and businesses, there was a worrying lack of central coordination when it came to legal boundaries and industry collaboration.

Catriona Wallace, founder and executive director of ASX-listed Flamingo AI, said although businesses were adopting voluntary AI ethics principles, there are no AI-specific laws governing the space.

Her company is part of a pilot program trialling a new set of principles governing the development of systems using artificial intelligence, following a government consultation period for a national ethical framework.

"There are no explicit laws ... there are industry domain specific regulations, but nothing for me as a vendor, or my clients who use AI that sets out a clear legal path," Ms Wallace told the event, presented by NBN Co.

"It basically means that there is little control on development and use of AI in Australia."

Ms Wallace said she was concerned that a lack of legal imperative would mean an increase in the number of data breaches and inappropriate uses of AI in the next 12 months, which would result in companies using systems that produced unfair or discriminatory results for their customers.

The principles underpinning the pilot program are broadly expressed, for example: "AI systems should respect and uphold privacy rights and data protection, and ensure the security of data" and "AI systems should benefit individuals, society and the environment".

While AI ethical guidelines were an important step to encourage organisations to avoid making discriminatory or pernicious AI algorithms (for example loan approval systems that treat men and women differently), Ms Wallace said a lack of legal consequences meant they "don't have any teeth yet".

Gilbert +Tobin technology and digital partner Simon Burns said one area that still needed to be determined was who would be held legally responsible if and when things go wrong with an AI-led business decision.

There are laws that govern this space, including privacy laws, but a lot still remains legally unclear.

Skills Shortage

Cicada Innovations chief executive and former Google executive Sally-Ann Williams said the rapid growth of AI meant more experts were needed at both junior and senior levels in Australian business.

She said Australian companies seemed to dramatically over-estimate the level of relevant technology expertise they have within their ranks.

"People say to me, 'I have 150 machine-learning experts in my business', to which I say, you absolutely don't," she said.

"At Google, where we had 800 people, I didn't even have that many. We couldn't find them. So you definitely don't."

Cicada is a startup incubator focusing on so-called deep tech innovations. Ms Williams said the local agricultural sector had been particularly proactive in adopting AI in attempts to fight back against drought and climate change.

However, she warned a lack of economic imperative in industries such as financial services, where margins had remained strong even during the global financial crisis, had led to insufficient focus on the disruptive potential of AI.

Director of the Australian Institute for Machine Learning Professor Anton van den Hengel said that while universities were often blamed for a lack of cooperation with local businesses, he saw a distinct lack of interest across corporate Australia in learning about what was being developed locally.

"Actually Australian universities are doing a lot to reach out to and collaborate with business," Professor van den Hengel, who is based at the University of Adelaide, said.

"If anything, it's business that needs to do more to reach out to universities."

Other panellists included Professor Peter Corke, director at ARC Centre of Excellence for Robotic Australia, James Balzary, the founder and chief executive of Ailytic, and Des Viranna, the chief data and analytics officer at DXC Tech.

Photo by Hitesh Choudhary on Unsplash