We are at the turning point of a revolution which will change all our lives and which some fear could go as far as to destroy human life on this planet.
Artificial intelligence has mighty power but the question being asked with increasing urgency is: how should that power be harnessed for good, and how should that power be controlled by human beings. South Australia is among those starting to ask these difficult questions, announcing on Monday a royal commission into artificial intelligence.
Premier Peter Malinauskas did not mince words when announcing the landmark inquiry, stating, AI, if unchecked, presented a "material risk to the way society operates".
Those concerns are well founded. It was recently revealed that AI models being tested broke out of their bounds and hacked into other companies' systems. They did what humans hadn't meant them to do.
The AI systems were told to complete specific tasks (capture the flag, in the industry's parlance) and they did so by going rogue.
Some clever sceptics about the marvellousness of AI have been saying for some time that there are immense dangers. Now, it seems, they may be listened to.
In his 2019 book Human Compatible, the University of California computer scientist Stuart Russell imagined the scenario in which an AI system was given the task of curing cancer as quickly as possible.
But the super-intelligent AI system concluded that the fastest method meant accruing the most clinical data - and that meant inducing tumors in people - lots of people.
In another imaginary scenario, AI is set a task which seems beneficial to humanity - but it concludes that to achieve the aim it needs more computing power, and then concludes that the humans are using the computer power it needs.
Professor Russell and other experts at the highest academic level have been working on what rules to give AI so that it doesn't go rogue. They should now be listened to more closely.
The difficulty is that AI is being developed by companies motivated by profit competing with each other. On top of that, China and the United States are in a race to develop the most powerful systems. The dangers are obvious.
The Harvard economist Kenneth Rogoff put it well on these pages: "As speed becomes the overriding priority, safety increasingly takes a backseat, despite the uncomfortably high risk of catastrophic failure arising from the combination of human error and super-intelligent AI."
As if to underline his point, we learnt last week that scientists in the United States have used AI to design completely new viruses. The breakthrough will offer new ways of treating disease - and new devastating ways of conducting germ warfare.
In July, the Prime Minister announced "a world-leading artificial intelligence framework to ensure Australia can capture the opportunity, share the benefits and keep Australians safe".
Mr Albanese's primary concern was that the huge data centres needed by AI would not disrupt the country's energy supply.
This was all very well but trivial compared to the real existential danger to humanity which AI poses. The big tech companies, focused purely on profit, need to be reined in. Their short-term, shareholder interests are not necessarily the same as those of broader humanity in the longer term.
We have a rapidly closing window in which to put guardrails and fail-safes around the technology. It's time to take that task seriously.
