Bruce Arthur: Could AI be a trap for humanity? The evidence seems to be piling up

Whatever a soul is, the brain is where it starts. I'd prefer not to outsource mine, writes Bruce Arthur.

A study from this summer, tracked over 26,000 Chinese students over two and a half years and showed homework scores improved, even as exam scores dropped 20 per cent within six months.

Bruce Arthur is a columnist for the Star. Follow him on Twitter: @bruce_arthur.

Sometimes my wife knows where we’re going and I don’t, so she directs me: left, then a right past the gas station, it’s the red building, there we go. Afterwards, if I have to find my way back to that place without her, I won’t remember how to do it. Because it was her brain that got us there, not mine.

That, for many, is the emerging experience of AI. Broadly, generative AI is sold as an easy button in this hard life. Summarize things for me! Find things I would otherwise have to search for! Paint a picture for me, make a movie. Or, in politics especially, write a speech.

Oh, and it might wipe out humanity, say many who are working on it. The writer David J. Roth offered the best interpretation of those comments, I think: when it feels like the world is coming apart, there is a certain childlike comfort in imaginary futures, even apocalyptic ones.

The removal of friction, of course, is a fundamental human goal. A project called Ordinary Abundance recently showed how far we have come in almost every way, using just the conveniences of a modern apartment. People wrote poems about washing machines, once.

More and more, though, AI appears to be a trap, and the studies are piling up. One, from MIT last year, called the effects of repeated ChatGPT use the accumulation of cognitive debt (they monitored brains during essay writing and measured the reduced activity for those using chatbots). Another, from this summer, tracked over 26,000 Chinese students over two and a half years and showed homework scores improved, even as exam scores dropped 20 per cent within six months.

MIT is one of the world’s finest universities, and is not opposed to technical innovation, but it recently released a study on the use of AI by its students, and it reflected the lived experiences of people throughout education. Fewer study groups; less use of office hours with professors; the increasing uselessness of take-home assignments and problem sets, with accompanying reductions in critical thinking, memory, confidence, and of course, mastery. The report did not call for a ban on the technology, but it did call AI-assisted study “the illusion of learning,” and warned overreliance could cause “cognitive surrender.”

But one study, from academics from multiple institutions, goes further still. It theorizes that the use of AI, and eventual dependence, is a cognitive virus. As it says, “As cognitive delegation becomes more widespread, the social environment that supports autonomous reasoning can weaken; as that environment weakens, delegation becomes still easier and more attractive … Beyond a critical point, the loss of autonomy becomes self-reinforcing and the population can move rapidly toward a state of much stronger cognitive off-loading and lower cognitive competence”

It states there is a difference between AI which is mental scaffolding vs. substitution, which is critical. The preprint also states that once cognitive devolution has occurred, achieving our previous state of cognition may not even be possible. Prevention, therefore, is preferred.

And this is before we get to how generative AI produces persistent hallucinations, or how human-made art remains so far superior but also necessary to the human soul, or how ChatGPT reportedly gave help to the Tumbler Ridge shooter earlier this year, or how AI-aided psychosis appears to be real.

AI companies, of course, don’t care. In his interview with the New York Times, Nvidia CEO Jensen Huang — a relative humanist, as AI magnates go — admitted skills like basic math could be lost.

Ezra Klein: But it must be some set of skills that matter.

Huang: Oh, yeah, yeah, yeah. But maybe not those. We’re going to discover new ones. Just maybe not those. There are a lot of skills that don’t matter. My first confession, I actually don’t know my address.

I can only remember about seven phone numbers, but an address seems like a river most people shouldn’t cross, and we need red lines about which core competencies matter. Studies have long shown the most beneficial teachers are the toughest ones, because friction — the hard process of learning, rather than simply pressing the button for an outcome — helps long-term outcomes. We know everything humanity has achieved has been the product of striving, and of the human soul. And we know the world is darkening and getting harder, and that humanity will need to be smarter, and more human, to avert the worst outcomes.

So perhaps outsourcing our brains in a race to grasp a new future is not a Luddite’s position, in the grand scheme. A CFL player and I once talked about concussions in football, and how scary it can be to lose your mind. He said a concussion isn’t an ankle injury, or an elbow, or a hip. It’s you.

Personally, I avoid AI, and have been using GPS less lately, when I can; it’s not much, but it’s something. Whatever a soul is, the brain is where it starts. I’d prefer not to outsource mine.

Error! Sorry, there was an error processing your request.

There was a problem with the recaptcha. Please try again.

You may unsubscribe at any time. By signing up, you agree to our terms of use and privacy policy. This site is protected by reCAPTCHA and the Google privacy policy and terms of service apply.

Opinion articles are based on the author’s interpretations and judgments of facts, data and events. More details

Bruce Arthur is a columnist for the Star. Follow him on Twitter: @bruce_arthur.