This article analyses the development of artificial intelligence (‘AI’) by the Australian Taxation Office (‘ATO’) which has, to date, focussed on increasing the likelihood of detecting taxpayer non-compliance. It finds that higher rates of detection should reduce tax crime and hence the tax gap. However, detection does not distinguish fraud and evasion from mistakes and errors. Combined with low tax morale, simplicity and literacy among Australian taxpayers, this is likely to reinforce a culture of fear, especially among individuals and small businesses. Fear-based systems undermine voluntary compliance. This article therefore argues that, without parallel investment in AI models targeted at improving tax morale, simplicity and literacy, the ATO’s detection models risk deteriorating voluntary compliance. This article recommends that the ATO should broaden its use of AI to address these issues and suggests a number of new AI-driven solutions which the ATO could implement. This analysis also provides useful insights for other public departments and global tax administrators implementing AI into their systems and services.
No takes yet. Share an insight, caveat, or question.
Michael Croft (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: