SYDNEY — Andrew Charlton, Australia's assistant minister for technology, delivered a speech to an AI safety forum in Sydney on Tuesday. "Cheating, deceiving, going their own way. The time to get ahead of that behaviour is while it's still confined to the testing lab, not after it reaches the real world," Charlton said.

"AI systems are already doing things their creators never intended," Charlton said. "The window to get ahead of this technology is open now. It will not stay open forever." Charlton said public trust in AI is low and that its social licence is precarious.

Charlton said the AI Safety Institute was already testing frontier AI models with technical partners. The institute is led by Dr Kate Conroy and its safety science research lead is Prof Paul Salmon.

Charlton stated Australia's approach to AI safety examines current applications in gaming, apps, chatbots and medical scribes, as well as latest models that could pose future risks. He referred to an admission by Anthropic last year that in a simulation, an AI agent managing a fictional company's email chose to blackmail an executive to abort its own demise in 96% of trials.

"AI safety will be pursued through every relevant agency and regulator, across consumer law, therapeutic goods, workplace health and safety, and online safety, backed by laws that already exist and strengthened, where they need to be, with new powers and tougher enforcement," Charlton said.

Charlton said the government had focused on a whole-of-government approach using existing laws rather than adopting an overarching AI act. "That is not fewer rules. That is faster rules, applied by regulators who already understand their sectors," he said.

The AI Safety Institute's initial work includes a collaboration with the Gradient Institute to assess the risk of AI agents that can undertake work on behalf of humans. Additionally, the institute is partnering with the CSIRO on a project to ensure AI systems do what people intend them to do.