Copyright Agency welcomes AI guardrails proposals
September 12, 2024
On 5 September, the Department of Industry, Science and Resources (DISR) released a paper setting out proposed mandatory guardrails for high-risk AI development in Australia.
The paper setting out the proposals is available here, the Department’s announcement is available here, and the Industry Minister’s announcement is here. The Department is seeking feedback by 4 October.
Copyright Agency welcomes the proposals as an important step towards regulating high-risk AI and addressing concerns in the Australia community about AI, and will be responding to the paper.
The proposals include requirements that AI developers:
- ensure that data is obtained legally
- disclose information about the sources, acquisition and processes of the data they have used
- address issues associated with Indigenous Cultural and Intellectual Property (ICIP)
The paper also refers to the intersections of these proposals with the work of the Copyright and Artificial Intelligence Reference Group (CAIRG) in the Attorney-General’s Department (AGD), of which Copyright Agency is a member. DISR is working closely with ADG and other Government departments on a coordinated whole-of-government approach to AI. Copyright Agency’s position is that the Copyright Act should not be amended, because it is already fit for purpose for ethical AI development in Australia (including by using licensed datasets) and new AI ‘exceptions’ would be damaging for Australians, but that the Government should introduce standalone legislation that compensates Australians for the use of their work in offshore training that underpins products and services available in Australia.
DISR is seeking feedback both on the proposals themselves, and how they would be implemented. It sets out three potential approaches:
- adapting existing regulatory frameworks to include the guardrails
- introducing framework legislation that will require other existing laws to be amended for the framework legislation to have effect
- a whole of economy approach – introducing a new cross-economy AI Act (the model adopted in the EU and in Canada).
Voluntary ai safety standard
On 5 September, the Government also released the Voluntary AI Safety Standard (VAISS): here. The VAISS includes the 10 guardrails that the government is proposing to make mandatory, with guidelines on application. The Industry Minister said that VAISS:
- provides practical guidance for businesses where their use is high risk, so they can start implementing best practice in the use of AI
- gives businesses certainty ahead of implementing mandatory guardrails
- in step with similar actions in other jurisdictions – including the EU, Japan, Singapore, the US – will be updated over time to conform with changes in best practice
- will help domestic businesses grow, attract investment and ensure Australians enjoy the rewards of AI while managing the risks
If you would like further information about the proposals and/or VAISS, please contact the Member Services team.