Globe and Mail Comment

Today’s Globe & Mail featured an editorial on AI risk, following an interview that the entire editorial board had with Geoffrey (“godfather of AI”) Hinton. Note: you’ll need a subscription to read it.

It’s a pretty good editorial (although it seems overly credulous in citing “medical breakthroughs” that amount to suggestions for new molecules to test, but that’s a quibble). They cover all the main points of AI Safety and don’t shy away from the difficult questions. Importantly, Hinton’s estimation of 10-20% probabilty of an extinction risk event (for humans) is put in front of us. Once again. Hinton’s “solution” – maternal AI – gets a shout out.

Prof. Hinton has an alternate vision, if only we could figure out how to do it. He proposes engraining AI with a maternal instinct, noting that parents have far more power than their children but are generally inclined to use that power to shield them, rather than to wield it against them.

Artificial intelligence that is smarter and more powerful than humans could easily destroy life on earth. But not if it is built to want to take care of those pitiful creatures who invented it, and then got left behind. If it is built to have empathy for humans.

Is this to be what we pin our hopes on? That the AI will look on us like a kindly mother and look after us? That seems like something that is both improbable and prone to pitfalls; assigning a superior intelligence a “mother” role just seems perilous.

We can do better.

William Leiss penned a commentary to the editorial, just to drive home the point that we don’t have to content ourselves with making observations about the probability of doom. Here is his comment:

The “godfather of AI,” who should be regarded as knowing what he is talking about, says that AI brings a 10-20% chance of human extinction. Pause for just a moment to ponder this statement. In a forthcoming book with Richard Smith, “From Tool to Actor,” we argue that this level of risk (for catastrophic outcomes) is so far above the norm for acceptable risk in Canada and elsewhere that it is off the risk charts. We know how to regulate technologies with risks of catastrophic outcomes, the best example of which is nuclear power plants. But our pathetic “AI for All” so-called strategy simply ignores all the most serious risks, even though someone like Dario Amodei, CEO of Anthropic, has clearly identified them. The Government of Canada has put its citizens in an intolerable situation and citizens must pressure their government to change course. There must be at least a global enforced pause in Frontier AI development until we are assured that new AI models will be safe prior to deployment (even the USA under Trump is moving in this direction).

William Leiss, OC, PhD, FRSC, Professor Emeritus, Queen’s University [email protected]

The Globe didn’t publish it.