There are a couple of features that stand out about the National AI Plan unveiled by the federal government this week. The first and most obvious is that it is not a plan at all.
Rather, it delivers a set of blue-sky aspirations or objectives, but without describing the steps to reach those objectives. Without detail or focus, this is a plan that smells a lot like wishful thinking.
The other thing that stands out is that it is wholly lacking in ambition for building Australian AI capability, and all of the wealth-creating economic activity that comes with that.
In relation to building foundational AI capability and creating Australian companies and Australian IP, there is essentially nothing to see here.
AI industry policy is all a bit of a snore in the mainstream, which is probably why this document is bereft of Australian industry ambition. And while Australia is snoozing, the global tech industry is arriving and staking out their ground.
This plan does not deliver the kind of national policy initiatives of a scale to meet the AI moment. In fact, it is so lacking in focus on Australian industry as an ‘AI maker’ that it is hard to believe the plan was written by our own Department of Industry. It has Treasury’s fingerprints all over it.
It perpetuates Australia’s role as a Howard era “sophisticated user” of technology, rather than as a creator, sending our wealth overseas as a result.
Australia’s chief scientist Professor Tony Haymet might be urging Australians to become ‘AI makers not just AI takers’ – but he won’t get any help through this plan.
There is no ambition for building Australia’s AI talent pipeline with the best and brightest.
When the Canadian government announced in its Budget last month $1.3 billion toward attracting 1,000 high-level international AI researchers – and training a pipeline of graduate AI researchers – you might think the Australian government might be thinking along similar lines.
In fact, UNSW Professor Toby Walsh had been advocating with the Kingston Group over the past two years for the Australian government to invest in attracting or training 2,000 PhDs in AI.
The price tag would be in the many hundreds of millions of dollars, which doesn’t seem too bad considering it would deliver a substantial proportion of global AI research talent to this country.
What we did with that talent would then be up to us. But at least we would have the option of building, building, building in Australia, rather than simply sourcing everything from overseas.
Build IP, create wealth. We need top-tier talent to build high value products and services that do not simply rest on foreign-owned tech.
It is worth noting that when this process was started a year ago, former minister Ed Husic had called for the development of a National AI Capability Plan. By the time it was delivered, it had been truncated to simply a National AI Plan. If we can agree that words have meaning, it’s not a great time to lose “capability”.
The plan says (under a heading of “what we’re doing next”) that the government is “agreeing collaborative arrangements with AI innovators” – by which they mean global AI companies.
Can you believe it? That’s the plan! In a nutshell, that’s the whole thing.
Here’s a tip. These global AI companies are extractive. They don’t want to collaborate with you. They want your energy, your land and your water. And they want to sell to you and your citizens.
Clearing a path for hyperscalers to set up here to deliver services is not the fulfillment of Australian ambition. It’s the fulfillment of the hyperscalers’ ambition.
We want those companies here, sure. But we need to be clear-eyed about what we can extract from them (and we must be hyper-vigilant about the energy systems to ensure they do not drive up power prices for our own industries and our own consumers, as it has elsewhere in the world.)
But that can’t be the extent of our ambition for AI.
We need a plan.
In the news this week:
Victoria to dismantle LaunchVic in cost-cutting drive
AI plan ‘a cop out’ on sovereign capability and energy
Victoria govt backs tech overhaul for public service
SERD findings must focus on business needs and outcomes
Australians must become ‘AI makers, not AI takers’, chief scientist warns
Robyn Denholm departs Tech Council prompting board refresh
Do you know more? Contact James Riley via Email.