Governments around the world are now recognising both the significant potential opportunities as well as risks associated with artificial intelligence (AI). As the Australian government seeks to regulate AI, it’s important that developments are for the benefit of the majority of Australians, and not just for the large technology companies that are creating AI models and products.
We recommend that:
- Australia establishes an AI Commissioner, who is able to take a whole-of-economy approach to regulate AI under a new Australian AI Act
- Details of ‘high-risk’ settings for AI are specified and clarified in a list, listing what these specific activities are, so that there is no confusion or uncertainty
- A similar definitional list is adopted for ‘medium-risk’ and ‘low-risk’ activities to minimise the potential of any activities being misattributed or miscategorised, and address the approach for other related AI issues such as privacy or misinformation.
- That the position for ex-post facto applications and liability of these proposals are clarified, given many significant AI models and products are already in market, and continuing to be updated at a rapid rate.
