Australia's AI Regulation: Lessons from Gene Technology Oversight

Australia's Current AI Regulation Approach

Australia has been striving to become a leader in "safe and responsible" artificial intelligence (AI) since 2019. The Department for Industry, Science and Resources introduced a voluntary framework based on eight AI ethics principles, including "human-centered values," "fairness," and "transparency and explainability" 1

1

2

2

. However, these voluntary principles lack enforcement power over organizations developing and deploying AI systems.

Recently, the Australian government acknowledged that "voluntary compliance [...] is no longer enough" and proposed "mandatory guardrails for AI in high-risk settings" 1

1

2

2

. Despite this shift, the core idea of self-regulation remains prevalent in the proposal.

Limitations of the Proposed AI Regulation

The current proposal for AI regulation in Australia has several shortcomings:

1. Self-assessment: AI developers are responsible for determining whether their AI system is high-risk 1

   1

   2

   2

   .
2. Lack of oversight: There are no provisions for external oversight, consequences for non-compliance, or mechanisms for refusal or redress 1

   1

   2

   2

   .
3. Minimal requirements: Companies only need to demonstrate internal processes that align with AI ethics principles 1

   1

   2

   2

   .

The Gene Technology Regulation Model

Australia has a successful regulatory model for gene technology that could be applied to AI. The Office of the Gene Technology Regulator, established in 2001, has become an exemplar of expert-informed, transparent regulation focused on a specific technology with far-reaching consequences 1

1

2

2

.

Key Features of the Gene Technology Regulator

1. Single-mission focus: The regulator's primary goal is to protect public health, safety, and the environment from risks posed by gene technology 1

   1

   2

   2

   .

2. Sophisticated decision-making structure: Risk assessments are informed by sound expertise and insulated from political influence and corporate lobbying 1

   1

   2

   2

   .

3. Continuous public input: The regulator integrates public input into its risk assessment process transparently and meaningfully 1

   1

   2

   2

   .

Applying Gene Technology Regulation Principles to AI

To improve AI regulation, Australia could adopt similar principles:

1. Clear mandate: Establish a single-mission regulatory body focused on protecting public interests in AI development and deployment 1

   1

   2

   2

   .

2. Expert-driven decision-making: Develop a national evidence base informed by cross-disciplinary scientific, socio-technical, and civil society expertise 1

   1

   2

   2

   .

3. Public involvement: Create meaningful opportunities for public input and consent in AI development and deployment 1

   1

   2

   2

   .

Lessons from Gene Technology Regulation

The success of gene technology regulation in Australia demonstrates that:

1. Innovation is not hindered by regulating new technologies until they prove safe for people and the environment 1

   1

   2

   2

   .
2. Comprehensive oversight can be maintained even when technology becomes widespread 1

   1

   2

   2

   .
3. Public safety and trust can be ensured through transparent and inclusive regulatory processes 1

   1

   2

   2

   .

By applying these lessons to AI regulation, Australia can create a more robust and effective framework for managing the risks and opportunities presented by artificial intelligence technologies.