PC issues AI warning
The Productivity Commission wants Australia to balance AI innovation with regulation.
Artificial intelligence (AI) is poised to redefine the landscape of productivity and economic growth in Australia.
The Productivity Commission's latest suite of research papers casts a light on this potential, urging a nuanced approach to governance that could see the nation harness benefits while mitigating risks.
The commission is calling for a cautious regulatory stance that leans on existing legal frameworks, suggesting that the rush to impose new AI-specific legislation could stifle innovation and place Australia at odds with global standards.
“Many of the potential harms that could be created by using AI are 'old wine in new bottles' – harms that we have previously encountered and that are adequately dealt with by existing laws and regulations,” the commission says, challenging the clamour for AI-specific controls.
Commissioner Stephen King, a Monash University economics professor, says Australia's robust regulatory environment is a strength.
“Before jumping to new AI-specific laws, we should examine existing regulations and better explain how they apply to the uses of AI,” King says.
The Productivity Commission acknowledges the global market's likely concentration in AI hardware, cloud services, and advanced machine learning models, areas where Australian policy may have limited sway.
Yet, it argues against the notion of a “perfect world” risk measure, advocating for realistic benchmarks, such as comparing self-driving vehicle algorithms to human drivers rather than an idealised zero-fatality world.
This balanced viewpoint contrasts sharply with public sentiment, where there is a strong desire for more stringent controls to protect against personal data abuse.
In a recent government survey, 90 per cent of respondents expressed a preference for tighter regulations.
The Productivity Commission is calling for a comprehensive national data strategy to facilitate economy-wide data sharing and foster the development of local datasets for AI learning.
Australia's position as one of the OECD's laggards in data availability, accessibility, and government support for data re-use underscores the urgency of such a strategy.
Using Australian data “would increase the relevance and grounding of AI applications to our needs,” the commission says.
The research presents a roadmap for Australia's AI journey, advocating for a balanced approach that leverages existing regulations while remaining open to future adjustments.
The experts say that by fostering an environment that encourages innovation and addresses public concerns, Australia could indeed ride the wave of AI innovation to new heights of productivity and economic prosperity.