When it comes to new technology, being first can come at a cost

This columnist recently had the good fortune of a visit to Greece.

Wonderful country. Rich in history. Ethos, pathos and logos by the bucket load. Which they invented of course.

Another discovery the Greeks are sometimes (debatably) credited with — thanks to the story of the Minoans at the Palace of Knossos in Crete in 1700 BCE — is flushing toilets.

The irony, as anyone who has spent time there knows, is that the Greek sewer system can't handle dunny paper because the pipes, at around 50mm, are too small. Aussie residential pipes are double that width.

The result is that everyone is told to put the paper in a nearby bin or risk setting off a hideous plumbing disaster.

During the visit, our algorithm-enabled Instagram feed filled with posts from tourists poking fun at this jolly reality.

Greece's predicament — of being vulnerable to a single foreign visitor's indifference to the basic laws of fluid dynamics — is an imperfect reminder that there can be a curse attached to being first in class.

If you value reliability, safety and flexibility, you might avoid adopting the initial iteration of a technology.

Being first can mean getting lumped with a bad outcome for a long time. Sometimes it's better to let others take the hit and learn from them. Flush out all the problems, so to speak. And put in bigger pipes.

This notion sits near the surface of the federal parliament's ongoing joint select committee inquiry into artificial intelligence, which is looking at both the opportunities and risks of the technology for Australia.

The Albanese government has imposed a tight timeline on the committee and is working with the Coalition on a quick outcome.

The committee has until the end of November to report, ahead of the introduction early next year of legislation to set standards for large data centre operators and settle the treatment of intellectual property and artist's rights. It aims to also shape responsible AI usage.

Such is the rush, that Labor and Coalition committee members have already issued a raft of recommendations this month that would require AI companies to submit to safety checks, reserve computing power for local use and provide security agencies with frontier models.

There are good arguments for not dawdling.

AI is expanding at a breathtaking pace around the world and there's a stunning wave of capital on offer.

"If we dither, if we do nothing, then of course that opportunity will sail by our country and our adversaries will be taking advantage of that opportunity at our expense," said committee deputy chair and Liberal Senator Jonno Duniam earlier this month.

According to Treasurer Jim Chalmers, Australia is one of the "world's top ten markets for data centre capacity".

Some $150 billion of AI infrastructure is slated for construction in the country over the next five years. In its submission to the committee, Treasury says global investment by hyperscale cloud computing providers will exceed $1 trillion this year.

So there's logic in ramping up efforts to lure those dollars here.

"We are a digital hub in the Indo-Pacific with the land, energy, skills and capital to make the most of the AI investment boom," Chalmers told a Citi economic conference in Sydney on Friday.

But there are also those who say it's better to hasten slowly.

The committee has been flooded with more than 500 submissions, many of them voicing deep community concern about where AI is headed. And at what cost.

If the federal government thinks it's on a winner by aggressively driving the country towards adoption and investment, it may have a big job ahead convincing voters that there are benefits.

There are many drivers of this scepticism, but one of them is the age-old fear of disruption. And it's always worse when there's a sense of control relinquished or stolen.

"Unlike any technology we've ever dealt with, the business model of AI is to generate mass unemployment," economist Richard Denniss told the committee a few days ago.

His argument is that the eye-wateringly high — if recently wobbly — stock prices of AI giants is "literally based on the assumption that hundreds of millions of people will lose their jobs".

Denniss challenged the committee to think through questions about how the claimed benefits of AI would be shared between Australians and foreign-owned AI firms; and what a wave of data centres means for electricity, water and construction costs, and inflation more broadly.

"Will the government be spending more money to equip public servants with AI tools, or will the government be funding AI rollout through job shedding?"

Arguments about AI are complex and rapidly changing, but the speed at which the government is pressing ahead will do little to ease a jittery community if voters feel they've never had a proper say in the matter.

"The only way the Department of Health can spend a billion dollars on AI is if it spends a billion dollars less on humans or if you guys increase their budget by a billion dollars," Denniss told the parliamentarians.

"All that money that Anthropic is betting on is coming from somewhere and that's going to have very real consequences for individuals, for communities, for the economy, for the way Commonwealth budgets are drawn up."

Prominent Egyptian-American economist and investor Mohamed El-Erian, writing alongside two colleagues in a piece published this week by Foreign Affairs, warns of the "narrow but feasible path to collective good" from AI.

El-Erian notes that AI is a major issue in American politics, driven by understandable anxieties.

"AI adoption risks repeating the mistakes of globalisation, which generated large and widely spread economic benefits but led to concentrated losses of jobs, growth, and industry in particular communities," they write.

"Government support in response was insufficient or arrived too late. Although the economy gained overall from free trade, the political bargain failed to give affected workers and communities a credible stake in the gains, even as individuals benefited as consumers from lower prices.

"The result was a backlash that contributed to the protectionism and fragmentation that continue to plague today's politics and society."

In its submission, Treasury is largely upbeat about AI and its potential, but there are jarring caveats.

AI will "increase the scale, speed and sophistication of information weaponisation and radicalisation".

"It is expected to lower barriers to harmful biological and chemical activity and simplify criminal and terrorist activity planning," the submission states.

Furthermore, according to the department, "women and girls currently experience disproportionate harm from AI-facilitated abuse, including deepfakes and targeted harassment".

The flip-side of this argument is that there's a case for governments to have access to AI for security and defence. Denniss stated that the use of AI by intelligence services is largely unrelated to the boom of data centre builds.

"There's plenty of technologies that the military use that I do not have access to.

"But this rush, this sort of generalised assertion that if we don't say yes to everything now, we will miss out is, I would suggest, unprecedented."

Senator David Pocock accused the government this week of being "starstruck" by US AI giants like OpenAI and Anthropic, and he called for a 5 per cent tax on data centre revenue.

"What our communities want to know is what will they do for us," Pocock said in a speech to the National Press Club on Wednesday. "How do data centres benefit us now and over the long term? Where are the protections for communities?

Jacob Greber is political editor of ABC's 7.30 program.