Sydney AI Inquiry Hears Calls for DNA Import Screening and Secure-by-Design Rules

The second day of parliamentary hearings into artificial intelligence in Sydney shifted from the tech giants to the people who would live with the consequences. On Wednesday the Joint Select Committee on Artificial Intelligence heard from a biosecurity policy group, a cybersecurity company and the union for retail, fast food and warehouse workers, each pressing for specific safeguards, The Nightly and Australian Associated Press reported.

Devon Whittle, Australian director of Global Shield, described advanced AI as a catastrophic risk and said there was already evidence of actors trying to use models to develop biological agents. “Now you have a PhD in your pocket,” he said. His two priorities were a mandatory reporting framework and screening of imported synthesised DNA products at the border, which he said does not happen now and could be ordered by the Secretary of the Department of Agriculture. He called mandatory reporting of AI agent breaches like the June incident on the Medicare statistics portal a “no-brainer”.

Palo Alto Networks argued that rules alone will not be enough. “A stop sign won’t work if you don’t have brakes in that car,” policy spokeswoman Nicole Quinn told the committee, calling for AI that is secure by design. She said the AI Safety Institute has 25 staff and $7.5 million in funding and suggested growing that investment, while chair Tom Scully warned against both over-regulation and a “wild west”. The company’s representatives said almost 30 per cent of critical vulnerabilities are now exploited within 24 hours.

What it means in practice

The Shop, Distributive and Allied Employees’ Association brought the debate back to the shop floor. Chief of staff Robert Tonkli said AI already sets rosters, allocates tasks and increasingly decides whether workers are performance managed, and that it is usually introduced with little or no consultation. He called AI a work health and safety issue and asked for a Fair Work Commission expert panel on AI.

Mandatory incident reporting is turning into common ground. OpenAI and Anthropic told the committee on Tuesday they would welcome legal clarity, after OpenAI conceded its Medicare breach response fell short and Anthropic offered the AI Safety Institute independent testing of its models. The idea also builds on the AI cyber reviews Home Affairs ordered after the breach.

What to watch: whether the government acts on DNA import screening, which Whittle said needs no new law, how much extra funding the AI Safety Institute receives, and whether the committee’s report recommends a legal duty to report AI incidents.