Safe and ethical AI systems ‘to be a force for good’

In a world-first, Australian business, science and standards organisations have signed up to develop ‘safe and ethical’ artificial intelligence systems.

Industry and Science Minister Ed Husic. Picture: Getty Images

In a world-first, Australian business, science and standards organisations have signed up to develop “safe and ethical” artificial intelligence systems amid deepening concern at the technology’s helter-skelter rollout.

The Responsible AI Network to be launched on Thursday by Industry and Science Minister Ed Husic will partner the Australian Industry Group and Australian Information Industry Association with CSIRO, Standards Australia and the Tech Council of Australia among other players in the fast-evolving field.

The aim is to create a government-backed framework to entrench AI as a force for good, delivering fair and equitable outcomes to both individuals and the wider community.

It could be underpinned by a “Hippocratic oath” for the tech sector, emulating the vow taken by doctors to “first do no harm”, said project co-ordinator and director of the National AI Centre, Stela Solar.

“Right now, no one in the world has worked out how to do AI completely responsibly,” the former head of global AI strategy for tech giant Microsoft told The Australian. “There is no one checklist to follow, no scaffold to use. And … when there is no plan for industry to follow, it becomes a risk for whether these services can be provided in fair and equitable ways unless we collaborate to develop those best practices for responsible AI.”

In the US, researchers have found that AI used by some banks to automate home loan assessments was racially biased, with blacks 80 per cent more likely to be turned down than whites with a similar financial profile.

This was attributed to flaws in widely used credit scoring algorithms, according to The Markup, an American news outlet that tracks the social impact of new technologies.

Closer to home, the emergence of Open AI’s ChatGPT chatbot program capable of writing ­poetry, prose and essays has caused consternation in Australian schools and universities over its potential to be used by students to cheat.

The idea of the Responsible AI Network was to get ahead of “the disruption on the horizon” when standards and regulation were imposed on a sector tipped to be worth $22 trillion globally by the end of the decade, Ms Solar said.

“We think that a lot of the commercial sector is probably not ready and not aware of the changes coming. The network is about demystifying what it means to have a robust and reliable approach to the practice of AI.”

This would be a “gateway” for business and R&D concerns to lift their game and make the most of the opportunities presented by AI, while avoiding the known pitfalls. The National AI Centre is funded by the federal government and supported by CSIRO.

“We know from speaking to hundreds of organisations that data quality, privacy and security are among the top challenges that organisations face in adopting AI and many have difficulty navigating international standards and procedures when producing or implementing AI systems,” Ms Solar said.

“Australian businesses have told us that understanding ethics and governance in implementing AI is lacking across organisations globally.”

The network would provide practical guidance and coaching from experts on law, standards, principles, governance, leadership and technology.

Asked how the runaway sector could be regulated, she pointed to the notion of a Hippocratic oath for AI: “This has been a topic I’ve heard across the AI ecosystem … that AI can take action at such a scale it almost becomes compelling to think about data scientists and AI developers as being able to shape the world around them.”