DSPANZ provided a submission to the Department of Industry, Science and Resources' consultation on Introducing Mandatory Guardrails for AI in High-Risk Settings on 4 October 2024.
In this submission we highlighted that many of our members agree that the government should take a more robust regulatory approach towards AI but that this approach should allow for international interoperability and innovation. Following our previous submission on safe and responsible AI, we noted that it was important to recognise the challenges that Digital Service Providers (DSPs) currently face when looking to leverage and invest in AI technologies.
With the proposals paper indicating that AI in employment processes can be considered high-risk, DSPANZ provided feedback on AI in employment software, recognising that our members offer software products and services to support employers and employees.
Our submission provided the following feedback on the proposed mandatory AI guardrails, including:
- We support the government following approaches taken by Canada and the European Union (EU) for defining high-risk AI systems to allow for global interoperability;
- DSPANZ prefers the principles-based approach to defining high-risk, acknowledging that a list-based approach may not keep pace with technological change and each of the underlying AI use cases;
- The mandatory guardrails must balance the need for transparency with enabling organisation to innovate and protect their intellectual property;
- DSPANZ supports options 2 or 3 for mandating the guardrails, but we recognise that the legislative approach could introduce conflicting requirements for DSPs who operationalise legislation driven by different government agencies.
Access a full copy of the submission here.