The Australian government has initiated an abrupt eight-week consultation aimed at determining whether certain "high-risk" artificial intelligence (AI) tools should be prohibited.
This move follows similar measures taken by the industry and other countries, including the United States, the European Union, and China, as they seek to comprehend and potentially mitigate the risks associated with the rapid development of AI.
As part of the consultation, the country's industry and science minister Ed Husic has released two key documents, that include:
- a discussion paper titled "Safe and Responsible AI in Australia" and
- a report on generative AI issued by the National Science and Technology Council.
Alongside these publications, the government has launched a consultation period, which will run until July 26.
#AI is already part of our lives. As the technology develops, we need to ensure it meets Australians’ expectations of responsible use. Be part of the @IndustryGovAu discussion, below. https://t.co/Gz11JCXlsG
— Australia's Chief Scientist (@ScienceChiefAu) June 1, 2023
Safe and responsible use of AI
The objective of the consultation is to gather feedback on how to promote the "safe and responsible use of AI."
It delves into various approaches, including the consideration of voluntary measures such as ethical frameworks, the potential need for specific regulations, or a combination of both approaches.
One specific query posed during the consultation directly asks whether any high-risk AI applications or technologies should be completely banned and what criteria should be utilised to identify such tools.
Risks
The discussion paper classifies AI in self-driving cars as "high risk", while a generative AI tool employed for tasks like creating medical patient records is considered "medium risk".
The paper highlights the positive applications of AI in fields such as medicine, engineering, and law, while also acknowledging its potential for harm.
This includes the misuse of AI in deepfake tools, the creation of fake news, and instances where AI bots have encouraged self-harm.
Furthermore, concerns regarding the bias of AI models and the generation of nonsensical or false information, referred to as "hallucinations," were also addressed.
Meanwhile, the National Science and Technology Council report emphasises that while Australia possesses advantageous AI capabilities in robotics and computer vision, its fundamental capacity in large language models and related areas is relatively weak.