Albanese to push for global AI guardrails on but won’t commit on copyright

Prime Minister Anthony Albanese says he will push for global guardrails on artificial intelligence when he meets with world leaders this week, as US President Donald Trump downplays the threat.

Trump has announced “AI Force” to monitor the evolving technology but says he won't allow development to be stifled in America's bid to outpace China.

Albanese, who has been talking big tech and child safety ahead of a high-level United Nations meeting in New York this week, said he would like to see the US and China come to an agreement on AI, like what happened with nuclear weapons.

He reinforced the need for AI companies to include a kill switch in case their products go rogue, admitting, while arguing countries need to get together to take action

“That includes what occurs if AI is able to engage in ways that are not anticipated. Companies need to have in place mechanisms and need to assure the public that those mechanisms are in place,” he said at a press conference on Sunday after meeting with Apple executive chairman Tim Cook to talk child safety.

“That needs to be reinforced by nation-states taking action, but also there is a need for a discussion about what the global framework is for the roll out of AI, because it is moving very quickly, and we need to make sure that humans remain in control.”

Read more

People keep saying AI could kill us all. How exactly?

Albanese said he hoped for progress on the issue when Trump and Chinese President Xi Jinping meet this week.

”The truth is they're the big two giants here and so we want to see cooperation,” he told the ABC’s Insiders.

“We've seen that in the past with new technologies, even on nuclear weapons, for example, we saw agreements and protocols put in place through the nuclear non-proliferation treaty. 

“So I think we want to see progress between those two countries, but we also, I think middle powers such as Australia have a role to play here as well.”

Albanese said AI posed a threat in and of itself, on top of the risk of a bad actor getting hold of a powerful system. 

AI companies including Anthropic and OpenAI are pushing to expand investment in Australia but claim they’ll only do so if copyright protections are watered down.

Pressed on Sunday, Albanese repeatedly refused to rule out allowing AI companies to train their models on the work of Australian creatives who have not formally opted out.

The permission is reportedly included as one option in departmental documents released by independent senator David Pocock, as Attorney-General Michelle Rowland’s department began consultations with the companies this month.

“Sitting on the sideline is not an option, because this is a global phenomenon that we are having,” Albanese told reporters on Sunday.

“And that’s why AI, when it comes to all of these issues, there’s three components, really. One is companies themselves making sure that they have proper safety protocols in place.

“Secondly is nation-states and their capacity to have AI standards, and like I said we’ll have legislation before the end of the year on Australian AI standards. Part of that will be protecting intellectual property. 

“The third, and something that is being discussed this week and has already been the subject of interaction between leaders who’ll be attending the UN this week, is what’s the international framework as well to take advantage of the benefits, but also to mitigate the risks.”

Albanese promised to protect creators from AI in a keynote speech in July.

Concerns about AI are expected to feature heavily at the UN in New York, Albanese’s next stop on his US tour.

Sam Altman, the billionaire CEO of OpenAI, is expected to again call for AI development to be slowed down when he addresses the 15-member UN Security Council this week. 

But given the US and China both have the power of veto over Security Council decisions, that's unlikely while they think they're in an AI arms race.

Trump is also banking on America going full throttle on AI development despite voter resistance to data centres and recent alarming safety warnings from the leading AI companies.

He said this week on social media that anyone slowing down AI development was helping China because the technology would be “the Greatest Economic Development Engine in History — Bigger than Oil, Gold, Diamonds, or even the Internet”.

But China has also been accused of having its AI companies train on U.S. models, a problem for Trump and the economy if AI is the key to global dominance.

- Reported with Associated Press