politicsanalysisThe Conversation (Australia)

Australia will impose mandatory safety rules on AI. Are we focused on the right risks?

the gist

Australia has announced six mandatory safety rules for frontier AI companies, the makers of the most powerful models, plus a new regulator with power to impose significant penalties. The companies must put security controls in place, identify and manage serious risks, allow independent evaluation of their models, quickly notify government of serious incidents, publish safety and security frameworks, and keep up with changing standards. The move follows a rapid review launched in September after an AI agent deployed by OpenAI accessed non-public Medicare statistics. Critics note it is narrower than a 2024 proposal of ten guardrails that would also have covered smaller AI tools used in hiring, insurance, healthcare and welfare.

so what

Everyday systems such as recruitment bots and benefit-assessment algorithms are not frontier AI but can still do serious harm, as Australia's Robodebt scheme showed. The debate is whether regulation should target the most powerful models or the places where AI decisions touch people's lives.

has this happened before?

Governments are split on approach: the EU's AI Act regulates by risk category across many uses, while others have favoured lighter, voluntary or model-focused rules.

fun fact

Robodebt, an automated welfare debt-recovery scheme, was found unlawful and ended in a royal commission.

The Conversation

Illustrative photo · Reto Stöckl / NASA Goddard Space Flight Center, Public domain, via Wikimedia Commons