Australia is moving toward stricter requirements for the use of AI in situations that carry higher risks, as well as for general-purpose AI models that can adapt to different uses. The government has proposed ten “guardrails” that would be mandatory for these scenarios, backed by an additional set of voluntary standards. The idea is to make sure that AI development and deployment happens safely, responsibly, and in a way that’s accountable to both regulators and the public.
Under this proposal, organizations that design, build, train, or modify AI (the “developers”) and those that supply or use AI (the “deployers”) would face new obligations. This is significant because companies that build AI systems often operate overseas, and it’s not yet clear how far-reaching Australian enforcement powers will be. For systems used in “high-risk” scenarios—ones that might pose serious harm to individuals or groups, or disrupt society—developers and deployers would need to establish transparent governance processes, monitor AI for potential biases, and give people enough information to know when AI is making decisions that affect them. Users must also have some way to challenge or appeal those decisions. The regulations would force developers and deployers alike to keep thorough records and be prepared to show compliance through conformity assessments, particularly before systems go live and whenever major changes are made.
Another key part of this plan applies to general-purpose AI. Although there’s talk of limiting the rules to the most “advanced” models, the current definition could still cover a wide range of tools, including well-known products like GPT and image-generation software. Because Australia is only now taking public feedback, the final shape of these rules isn’t certain. However, the government is encouraging businesses to align with voluntary AI safety standards immediately. These standards mirror most of the mandatory guardrails and clearly signal that regulators, investors, and consumers are watching AI risks more closely than ever. If your organization develops or deploys AI, now is the time to review internal policies, raise awareness of AI governance among employees, and prepare for the possibility that these proposals could become law as soon as 2025.
For our clients, these proposals highlight the importance of staying ahead of AI regulations and best practices. It’s crucial to have a solid framework in place to manage risks, ensure transparency, and maintain accurate records. By taking a proactive approach—whether you’re building AI tools or using them in day-to-day operations—you’ll not only protect your organization and your customers, but also position yourself as a responsible leader in a rapidly evolving field.
Our team is here to help you navigate these requirements and tailor an AI governance strategy that fits your unique needs.
