The AI Symposium continues. So far Democratic Renewal Correspondent Ashok Panikkar has written on thinking and Political Economy Correspondent Oded Faran has written on the jobs market. I then joined the party by talking about human judgment and AI bias in an era of The Great Genocide Blood Libel. In the spirit of democratic debate, Ashok has been invited back to give his response.
His work has produced the irony of ironies. Ashok used an AI model to help him with the final edit. He commented, in a now deleted paragraph, that he may have put a human editor out of a job. He didn't. I cut swathes of his original paragraphs because as we all know AI editors feel that the more words, diversions and verbosity the better. With my audience in mind, I used my judgment as a human editor. So perhaps Ashok's concerns argued below are unfounded. The key is finding the right human editor and here at The World's Intelligent Journal he found one. - Daniel Clarke-Serret (human editor, Guerre and Shalom)
The AI symposium to date:
“The Things We Forgot to Count” by Ashok Panikkar
On AI, human judgment and the invisible costs of progress
Oded J.K. Faran recently wrote a thoughtful response to my essay in Guerre and Shalom, SURE WE CAN. BUT SHOULD WE?, entitled AI Is Not Making Expertise Obsolete.
I enjoyed reading it, particularly the parts where he seemed to be arguing with himself. We could do with a little more of that these days.
I also appreciated his willingness to follow the evidence even when it complicated his original argument. In a world where people routinely announce their conclusions before examining the problem, that’s rather refreshing.
And I agree with much of what he says.
AI exposure doesn’t automatically mean job substitution, nor does substitution necessarily mean the disappearance of an occupation. And even if an occupation disappears, it doesn’t follow that we’re heading for mass unemployment. Fair enough.
But as I read his essay, being neither an accountant nor an economist, I kept returning to a nagging question. What if mass unemployment isn’t what we should be most worried about?
I suspect our disagreement has less to do with artificial intelligence than with what we consider worth worrying about in the first place.
A Job Is Not Just a Collection of Tasks
Oded describes a job as a collection of tasks held together by an employment contract, a desk, and a good deal of habit. But surely a job is more than that?
For most people, work is one of the places where they discover what they’re good at, what they’re terrible at, what other people expect of them, and what happens when they fail to deliver. They learn to deal with difficult colleagues, unreasonable bosses, and customers who don’t know what they want. They make mistakes, acquire habits, develop competence, and occasionally discover that they’re capable of rather more than they imagined.
Some even find meaning in their work. I know I do.
When we automate a job, or parts of it, we’re not necessarily just removing tasks. We may also be removing experiences that help people become competent adults and, dare I say it, better human beings. This becomes particularly important when we discuss the disappearance of apprenticeships for junior professionals. As I argued in my original essay, the boring grunt work was how juniors became seniors.
When I was in design school, long before CorelDRAW and Photoshop, we spent days, months, and even years learning to work with pencils, charcoal, paints, and brushes. We sweated through the Ahmedabad heat, struggling with anatomy, perspective, proportion, light, and shadow. We drew the same things over and over again, often with rather disappointing results.
But something else was happening while we were busy learning to draw and paint. We were learning to see. To notice things that had previously escaped us and recognize relationships between shapes, spaces, and proportions. We were developing an eye for what worked, what didn’t, and why.
We weren’t simply acquiring technical skills. We were developing design sensibilities, a way of seeing and making sense of the world.
I wonder how much of that would have happened if someone had handed us a machine that could produce a beautiful drawing in three seconds.
The young lawyer who spent hours examining documents wasn’t simply producing something a client could be billed for. She was learning to recognize what mattered. The junior analyst struggling with a spreadsheet was learning that numbers have a nasty habit of concealing assumptions. The young journalist chasing a story was learning that people don’t always tell the truth, even when they sound terribly convincing.
Much of this learning was incidental. Nobody designed it as a training program. It happened because the work had to be done.
Now the work may no longer need to be done by a human being. One may call this a potential training crisis. We might discover that we’ve become very good at producing professional-looking work while becoming rather less good at producing professionals.
However, I think the consequences could be even more serious. Today’s training crisis could become tomorrow’s expertise crisis. And eventually, without that finely honed expertise, we may find ourselves facing a crisis of judgment.
Nobody Asked the Horse
I particularly enjoyed Oded’s discussion of the horse and the teamster.
