Albanese’s AI speech was a good start. Now Australia must confront these bigger questions

Every new technology has a tipping point, when it speeds towards ubiquity, changing the world as we once knew it. In the past six months, artificial intelligence, in its various forms, has crossed that threshold.

You can measure it in the billions of dollars being invested; the hope-tinged-with-fear decrees from presidents, prime ministers, Nobel laureates and even the pope.

It promises to be the most transformative technological change since the introduction of electricity, touching everyone and everything as “AI companies monetize the collective knowledge of the human race”.

It will challenge centuries-old notions of national sovereignty and human autonomy in ways not seen since imperial colonisation was at its peak.

Will Australia become, as the John Curtin Research Centre asked in its May review of AI, a “cloud hostage” – physically hosted, legally governed and operationally controlled from abroad? For most people, it still feels a bit abstract.

What does it mean when Demis Hassabis, the Nobel prize-winning founder of Google DeepMind, says it’s like the Industrial Revolution times 10, at 10 times the speed? Or Dario Amodi, head of Anthropic, declares half of all entry-level jobs will disappear in a few years?

Disruption, chaos, exploitation, mass unemployment and loss of sovereignty, or opportunity and transformation as once impossible problems are solved.

Last week, before Anthony Albanese delivered his landmark speech, a series of personal encounters brought it home to me that AI is no longer a hallucinating plaything.

A colleague explained how her workforce now included more AI agents than human beings – agents with names and personalities who sent emails and responded to questions, but apart from tech support, didn’t need to be paid. Agents, who with the right prompts and supervision, could analyse years of data in a flash, and advise humans on what to say in negotiations. They were so good that before hiring a human, the threshold question in her business was: could an AI agent do it? In many cases the answer was yes – and better.

A retired academic had found that Anthropic Fable was an extraordinarily efficient research assistant. He briefed it to distill and compare four recent Australian government-commissioned reports about racism and why so little had been done to implement the findings.

AI suggested that the answer was the uneven distribution of power. Those with cultural and economic power were more successful – so when tackling racism, the disadvantage of the least powerful remained entrenched.

Now there’s something to remember.

The prime minister’s speech on Wednesday showed how seriously the Australian government is taking this. Albanese spoke clearly and cogently, with more than a touch of Australian exceptionalism.

He outlined processes that, if properly implemented, could eliminate the worst mistakes of the past.

Standing in front of a display of Australian native flowers, he talked about sovereignty and the need to ensure that the benefits of AI investment did not stop with the building of datacentres, to add value beyond the land grab.

“We cannot settle for a short-term boom in capital expenditure and construction; we must create a new generation of good secure jobs,” he said.

Most attention has been focused on building the datacentres needed to power AI and on the protection of intellectual property and copyright. These crucial questions must not be dodged.

But there are bigger questions that have not been asked: is Australia prepared to take advantage of the technology to develop its sovereign data capacity and unique areas of expertise? Can it protect its sovereign security in defence, infrastructure and communications? Will it be able to ensure that the output improves the quality of life and public services for all its citizens and residents? Are there plans for public AI infrastructure so that local researchers and innovators can also reach for the stars?

Investment by international companies is important, but if we don’t want to repeat the mistakes of the past, it’s not sufficient.

Over recent months, both Canada and the UK have announced major multibillion-dollar investments in strategic AI capacity. Public money is being used to help fund local ventures that protect data sovereignty and provide scope for development by more than the big five transnational corporations.

Albanese skated over the debates in his scene-setting speech. But they are alive around the world. The moral challenges of AI, described so graphically by Pope Leo, need serious attention. As do the profound ethical challenges that are leading AI workers to join unions or resign as their companies’ owners cut questionable deals with the military-surveillance complex. Remarkably, there is even serious discussion in the US, of all places, about developing sovereign wealth funds to help those whose jobs disappear, possibly funded by government taking equity in trillion-dollar AI corporations.

These structural questions are being flagged at the edges of the policy debate here, but need to be taken seriously. As Albanese said: “If we hang back, or stand still, this will run right over us.”

It will be a busy six months to turn this vision into legislation and reality. The triteness of Angus Taylor’s response, that the only jobs that would be created were in the prime minister’s office, missed the weight of the moment and the huge risks of getting it wrong.

Julianne Schultz is the author of The Idea of Australia