HTI submission to the Productivity Commission Interim Report: Harnessing data and digital technology

In his address to the National Tech Summit, Minister for Industry and Innovation, the Hon Tim Ayres MP stated that ‘Australia needs to commit early and sincerely to sharing the benefits of AI among everyone – not just AI firms or big corporate players, but everyone.’  

To realise this vision, Australia must take a proactive regulatory response to AI: one that ensures that productivity gains associated with AI adoption flows to individuals, and that AI harms are adequately addressed. Without this, we risk missing out on the many benefits AI has to offer.

HTI’s submission, in response to the Productivity Commission’s Interim Report, Harnessing data and digital technology, proposes that Australia adopt a balanced approach to regulation for AI, aligning with Australian values and community expectations. Our key proposals are set out below.

Getting the balance right

The AI boom is fuelled, at least in part, by optimistic modelling on what increased AI adoption could bring to the GDP and the broader Australian economy. Australia’s regulatory and policy response to AI should be founded on a realistic assessment of the opportunities and risks likely to flow from increased AI adoption. The modelling for upside and downside risks should be equally rigorous.

While the economic upside for Australia could be profound, good public policy also must take account of the downside risks. For example, AI is fuelling an increase in online scams, costing Australians $2 billion in 2024. Similarly, the experience of Robodebt is a reminder that, where automation is poorly implemented, the economic and human costs can be devastating, with taxpayers spending over $2.4 billion on remediation related to failures in the Robodebt scheme.

If we want AI’s productivity potential to benefit people, we also need to ensure that people and communities are at the centre of our response to AI. Otherwise we risk a future where a handful of large corporations capture most of the productivity gains, while Australian workers, communities and small businesses miss out, and bear a disproportionate share in the risks.

An Australian approach to regulation for AI 

Treasurer the Hon Jim Chalmers MP has called for a ‘middle course’ – one that pays proportionate attention both to AI opportunities and harms. HTI’s submission aims to give life to this middle course by proposing an approach to public policy and regulation for AI that rests on five key elements:

1. Focused regulatory gap analysis: The main risks associated with AI are already well understood, and so the Government should undertake a targeted regulatory gap analysis, focusing on these areas of greatest risk.
2. Preference for technology-neutral reform: Australian governments of all political persuasions have always preferred technology-neutral laws, because they tend to promote positive innovation. This should remain Australia’s preferred regulatory approach for AI, with technology-specific laws adopted only where a technology-neutral approach would be inadequate.
3. Prioritise existing reform proposals: There are a number of law reform recommendations that have been carefully devised and consulted on, which have been left sitting on the shelf. The Government should urgently implement these proposals, including the Attorney-General’s Department review of the Privacy Act, and the Robodebt Royal Commission’s recommended framework for automated decision-making by government.
4. Framework legislation: Australia needs flexible legislation to address high-risk AI. It should apply across the economy, and provide clarity and consistency. Framework legislation would provide a default set of rules, that could be departed from in other legislation to avoid unnecessary duplication with existing laws. It would set out a nuanced process to assess AI risk, and remain applicable to new AI technologies.
5. Regulatory enforcement: Laws are only effective if they are properly enforced. Regulators need the right resources, powers and capability to keep pace with AI – which requires active government support.

Good regulation for AI improves the prospect of achieving significant productivity uplift, by providing regulatory certainty for business, making AI products and services more trustworthy for consumers, and supporting the workforce through a major technological transformation. A failure by Australia to adopt a balanced and effective regulatory approach for would invite the worst of all worlds: increased AI adoption would not drive significant productivity growth, yet harms associated with AI would not be appropriately addressed.

Strong privacy protections must remain a priority  

The Productivity Commission correctly recognises the limitations in relying on consent in today's digital world, especially where consent is rarely genuine, free and informed. Indeed, this problem was at the heart of many of the Attorney-General’s Department recommendations for reform of the Privacy Act, which were committed to by the Albanese Government in 2023.

Those reforms were the product of several years of detailed consultation across all sectors of the Australian economy and civil society, and focus on measures such as the introduction of a ‘fair and reasonable’ test for how companies deal with personal information. The Productivity Commission has proposed a new ‘alternative compliance’ pathway, which is untested and uncertain. It would appear to weaken protections for individuals, contrary to community expectations that the Government act to protect people’s human right to privacy. HTI urges that the Productivity Commission support the Attorney-General’s Department’s reforms for privacy law.