The treasurer is telling us to stay calm, but this could be the time to panic

Sydney Morning Herald

9 hours ago

'AI may be the most transformative technology in human history,' wrote the treasurer this week.
It's the rarest kind of statement: at once emphatically grand and altogether too modest. Grand, because it has AI outstripping, say, the wheel, electricity, or the internet. Modest, because it puts AI on the same continuum as all this, as though its difference is merely one of scale and speed, rather than something more fundamental. In this way, Jim Chalmers echoes the spirit of the Productivity Commission's latest interim report, also released this week, examining how AI will change our economy.
It would be fair to say the report is optimistic, seduced by AI's promise of significantly increased productivity. To that end, it warns the government against overregulating AI, on the basis that this would slow down its productive march.
Chalmers' version of the same idea has government regulating 'as much as necessary to protect Australians, but as little as possible to encourage innovation'. All of which presumes some kind of clear-eyed assessment of AI's risks. And it's hereabouts that all this optimism should give us pause.
Not because either the government or the commission deny there are serious risks. But because they characterise these risks – much as they characterise AI – as mere extensions of previous experience. Indeed, the commission's report could hardly be clearer on this point. After running through a series of potential problems – including some serious ones such as AI making mistakes in high-risk situations like healthcare or law enforcement – it concludes there is ultimately nothing new to see here: 'AI can exacerbate existing risk of harm but does not create wholly new risks where none existed before'.
Loading
Meanwhile, Chalmers acknowledges the possibility of significant unemployment, but believes it will not be widespread or structural. To this end, he makes the observation that while technological developments always eliminate jobs, they create more than they destroy. 'We've seen this play out before,' he affirms.
But this is more an assumption than an argument. It assumes that all technological advancement is some single, undifferentiated phenomenon, such that its history broadly repeats. But this is something the Albanese government must know not to be true. It is, after all, implementing a ban on social media platforms for children under 16, a belated response to a damaging technology we spent years assuming would be as benign as, say, television.
Now we seem to be assuming similarly AI will neatly fit into a benign pattern. That assumption only holds to the extent AI is analogous with most of what has come before. And in the circumstances, we'd be wise to examine it far more rigorously before settling on it because there are good reasons to suppose it is a different species altogether, for which history is a poor guide.