Andrew Hastie told a parliamentary inquiry that Microsoft Copilot suggested he reply "congratulations!" to a man with a terminal illness who had written that he planned to end his life with voluntary assisted dying.
The same exchange also produced suggestions of "great to hear from you" and "that is wonderful news!". Hastie used the example to argue for AI run by Australians, not software governed from the US.
Andrew Hastie at the inquiry
He raised the episode on Friday before the Joint Select Committee on Artificial Intelligence. The example was not presented as a hypothetical. It was a real response generated in a sensitive conversation where tone mattered more than speed.
For ordinary users, the practical issue is simple. An AI assistant can draft a reply that sounds polished while missing the human meaning of the message it is answering. In a health-related exchange, that can turn a drafting tool into a liability.
Abigail Bradshaw on cyber AI
Abigail Bradshaw told the inquiry her agents needed advanced AI every day to fend off "highly capable malicious cyber actors". She said Australia’s top cyber-intelligence agency is "entirely reliant" on US-housed processing and would lose capability "dramatically" if cut off in a conflict.
That leaves the policy debate in an awkward place. The inquiry heard a disturbing Copilot example at the same time as it heard that cyber agencies need access to the same class of tools to defend Australia and the army.
Chris Crozier and local control
Chris Crozier said the government was already trying to keep its tech needs local. He said it was contracting Google to build interlinked datacentres spread across the country and disconnected from the US.
He added: "Somebody can’t wake up and have a bad day and decide they’re going to turn it off." He also said: "We do not want to be using AI tool sets that are located offshore … we want the compute, we want the application, we want the data, we want the decision making, all based in Australia."
The Office of AI said bringing firms under Australian law would make it easier for local law enforcement to combat threats on the platforms. The government is consulting on standards that would force AI companies to report issues and breaches and potentially reserve some computational power for local research, but the source does not set out a final rule or deadline.
