We’re spending bucket loads on the thing that could kill us. Here’s why

After years of hearing about sluggish business investment – and how it’s a big reason we struggle to lift our living standards – you would think the recent surge would be a good thing.

And in many ways, it is. The most recent “national accounts” – a set of data gathered and shared every three months by the Australian Bureau of Statistics – showed business investment was the largest contributor to economic growth in the three months to March.

Business investment: spending by private businesses on things such as machinery and equipment, is important for growing the economy but also for increasing how much we all share in that growth.

Generally, when businesses invest in things that help us do our work more easily, it means we use fewer of our scarce resources (whether that’s ingredients, material, or time we spend working) to supply goods and services. That helps us become better off on average because we can have more of what we want and need.

Over the March quarter, private investment grew 3.6 per cent and the economy grew 0.3 per cent. The biggest driver? Business investment in machinery and equipment (up 16 per cent), largely made up of data centres across New South Wales and Victoria.

These are the engine room for modern websites, cloud applications and AI platforms, and as artificial intelligence has exploded in popularity, we’ve needed more of them to meet demand.

But data centres also come at a cost.

Most people, for example, know about the huge amount of water and energy guzzled up by these facilities. By 2030, the Australian Energy Market Operator reckons data centres will need as much electricity as all the households in Victoria.

Earlier this week, assistant minister for science and technology, Dr Andrew Charlton, said in a speech that data centres would be forced to wind down power use at peak times to prevent blackouts or strain on the grid, and that their owners would be required to build enough energy supply to power their operations (which could also then be accessed by households when those data centres aren’t in full force).

But as the government’s other economics expert, assistant minister for productivity, competition and Treasury – Dr Andrew Leigh – pointed out in a speech last month, there are even more consequential risks we aren’t paying enough attention to.

As economists – and students of economics – many of us are familiar with concepts such as economic growth: what it’s made up of, what drives it, and what might cause a slowdown.

The biggest part of the economy, by far, is households: their spending makes up more than half of Australia’s gross domestic product. While they weren’t the biggest driver of growth in the most recent quarter, consumption by households contributed 0.3 percentage points to growth in the three months to March, meaning they were the difference between an economy with sluggish but positive growth or no growth at all.

While higher interest rates dampened households’ appetite to splash out on “nice to have” things such as restaurant meals and alcohol, Australians did spend nearly 12 per cent more on essentials such as electricity, gas and other fuels (not because they suddenly needed more energy, but because of the end of a series of electricity rebates).

The role of government spending, meanwhile, took a back seat as state governments stopped spending on things such as electricity rebates.

While investment spending by government, businesses and individuals contributes to the country’s GDP, the amount this adds to economic growth depends partly on whether that money is spent on imported goods or those made domestically.

Why? Because the equation for calculating gross domestic product includes “net trade”: exports minus imports. When we import goods and services, they get subtracted in the formula because they haven’t been produced in Australia.

So, a lot of that growth in investment spending on imported equipment and machinery for data centres, for example, doesn’t actually boost our economic growth by as much as we might think at first glance. Sure, business investment might jump, but some of that effect is offset by the fact it’s imported from overseas.

Now, that’s not the thing that might kill us. Trade is generally a good thing that helps us swap the things we’re better than other countries at making (or doing) with things that they are better at pumping out: a win-win.

The big risk is linked to what our big jump in investment spending on AI tells us about our priorities, and where those priorities might lead us.

As Leigh points out, the earth has gone through extinction several times in the past: ice ages wiped out entire species and the dinosaurs were probably wiped out by a giant asteroid.

But today, the biggest risks to human existence are actually made by us: nuclear weapons, climate change, and deadly diseases engineered by people. Increasingly, AI has become a threat, too – and not necessarily because it’s evil.

We might think that as long as we give AI clear guidelines as to what we want (and don’t want), we’ll be okay.

But just because we give AI clear instructions, it doesn’t mean it will get them right, or follow them in a way that actually aligns with our interests. An AI that has been given a specific task, for example, might, in its quest to complete that request, trample over everything else that we failed to consider.

The concern, as Leigh puts it, is not simply that a model makes mistakes. “It is that a highly capable model may pursue the wrong goal in a clever, coherent, persistent, and hard-to-reverse way,” he says.

Even an AI system with the fairly innocent goal of making paperclips could, in its efforts to maximise its objective, find it useful to hoover up resources, resist shutdown, and manipulate its environment in ways that might be destructive to our other interests.

And given we’re already prone to making mistakes, the possibility of us unintentionally (or intentionally) creating an AI with misaligned goals is more likely than us getting it completely right.

The worst outcome, from a human-centred perspective, is that we get completely wiped out.

The reason economists don’t often think about disastrous scenarios like this is that standard tools of economics aren’t great at considering risks that are hard to put a price on, outcomes that can’t be compared against a familiar range, or low-probability events of enormous scale (especially when the probabilities themselves are uncertain).

Yet, the reason we’re not putting the brakes on AI development can be explained using economics.

There’s the free-rider problem, for example. If even one country took a step back and strengthened guardrails around AI, every other country would benefit from the boost in safety. But it’s one of those situations where everyone hopes someone else will pay the price for safer AI practices, and no one wants to pay for something that someone else will benefit from.

Meanwhile, because AI also comes with potentially huge profits and advantages, countries have more to gain from investing heavily in (potentially destructive) AI than in technologies that reduce the odds of catastrophe. The immediate rewards on offer for charging ahead with developing AI are high, while the immediate, monetary rewards of developing technology to contain it are fairly low.

And when it comes to governments, even though preventing our extinction seems a worthwhile cause, investing in things like AI guardrails and monitoring don’t benefit them much. Why? Because the benefit – averting disaster – is invisible and hard to prove: better to wait for a disaster to happen and come to rescue once the threats are visible.

Finally, it’s not something most of us probably think about too deeply because the individual benefit we get from learning deeply about existential risk is tiny: individually, there’s not much any of us can do about big issues such as AI.

But at the risk of sounding a little too apocalyptic, as we invest more into the technology, it’s worth us trying to put more thought and money behind not just boosting prosperity and economic growth, but about making sure we have an economy to talk about – and exist in – at all.

The Market Recap newsletter is a wrap of the day’s trading. Get it each weekday afternoon.