It is a welcome change to see Australia join with a global alliance of 22 countries to acknowledge the imminent threat of runaway AI.
The world cannot, and must not, stand still while the United States and China recklessly push frontier AI development ahead of our ability to limit and safeguard this technology.
Today’s statement now needs to be backed up by concerted action from Labor. It means seriously funding Labor’s tokenistic AI Safety Institute and following up with urgent, and binding, global agreements on public interest safeguards.
Greens AI Spokesperson Senator David Shoebridge said:
“This is a welcome change in tone from the Prime Minister, but it’s years behind the community’s concerns and it only comes after global tech bros voiced concerns about AI’s safety.
“The work to protect Australians from runaway AI needs to be more than a New York talking point.
“This statement needs to be backed up with funding a serious AI safety institute and concerted local and global action to agree on binding AI safeguards.
“For all the talk there is still not a single legislative response from the Albanese labor government on AI. Not one.
“For the past four years the government has sat on its hands, actively pushing back against calls for action from the community and the Greens by saying existing regulations were sufficient to deal with AI.
“Even where some action has been taken, it lacks real substance. The $7.5 million annual funding to the AI Safety Institute is so far short of the mark that it is embarrassing.
“Just one week ago Labor joined with the Coalition to vote down an urgent Greens motion calling for exactly what happened today, for Australia to join with other middle powers and start developing global AI controls.
“Global action may well put Australia on a path of conflict with the US and this will be a challenge, but failing to act would be an act of gross negligence. Now is not a moment to be timid given the scale of the emerging threat.
The scale of the threat from AI was also highlighted in the findings of an Independent UN Panel on AI that released its findings today – AI Agents, Misalignment and the Risk of Losing Human Control: Evidence from the OpenAIHugging Face Incident
