AI is being widely touted as a solution for productivity issues, but the productivity benefits of AI are not clear or guaranteed. Studies demonstrate that the path towards productivity using AI remain uneven and is subject to complex political, social and industrial factors. There are a raft of negative externalities currently being ignored in the AI productivity debate, such as harms to workers, copyright and privacy breaches, and new issues created by AI, such as the AI efficiency trap, cognitive debt, and inequality. AI is likely to be a transformative general technology which will influence most of Australian society, so its macro impact, its benefits and harms cannot be filtered through a productivity lens alone. There are other considerations around security, sovereignty, competition, privacy, human rights and other significant areas which require a national, whole of government approach.
We recommend therefore, that the inquiry consider:
1) Developing more research and studies which take into account productivity downsides, including existing studies, real-world cases and experiments which provide compelling counter-claims against perceived AI productivity benefits and provide a more holistic and realistic picture of the productivity impacts of AI.
2) That the rights and wellbeing of workers cannot be ignored as negative externalities, and that the copyright and intellectual property of Australians, and the harms and displacement on Australian workers and industries must be factored in.
3) That an AI Commissioner or central governance body be developed to account for the complexity and wide-ranging impact of AI across all areas of Australian society.