This week, a colleague received a spam email from an AI company—I won’t advertise it—that opened as follows:
“[Company name] is working with a number of Universities to create digital AI twins of professors that enhance asynchronous learning through always-available, personalized support.
“We’re seeking technical feedback from professionals in your field to help us refine the product.”
I don’t know if it was a prank; I hope it was, but even if it was, it’s plausible. It seems like the logical endpoint of AI and bots: Create AI versions of people and you don’t need as many people! Heck, after a while, the twins could replace the originals. The financial argument makes itself, at least in the short term, at least as long as you’ve never heard of the labor theory of value or have any concept of aggregate demand.
But I also recognized it as the plot of the 1981 movie Looker, with Albert Finney and Susan Dey, written and directed by Michael Crichton.
To be clear, Looker is not a very good movie. I remember watching it on a local UHF station on a Sunday afternoon a few years after it came out. Other than the plot, which somehow stuck with me, the most memorable part of it was the musical score.
The plot revolves around studios (I think) that had been trying to optimize the beauty of actresses in commercials and movies through plastic surgery. At some point, the studios figured out that they can instead create digital twins of the actresses and optimize those, thereby saving money and trouble. Of course, as long as the actual actresses were still running around, there was a risk that they’d do things to diminish the appeal of the digital twins. Accordingly, the baddies closed the loop by systematically murdering the actresses after they’ve been copied, thereby ensuring total control over their images. The rest of the movie became a murder mystery, with plastic surgeon Albert Finney piecing the conspiracy together and trying to nab the baddies.
For all of its flaws, Looker at least portrayed the manufacturers of digital twins as the bad guys. The sexism of the premise (and the camerawork) wasn’t exactly subtle, but there was a discernible (if flawed) strain of humanism in there. We were meant to assume that humans are more valuable than their digital twins.
I’d like to hold on to that, at least.
When I reflect on classes that I took in college and grad school, part of what I remember is the professors themselves. They were full, present humans, with the quirks, strengths and flaws implied by that. Some were funny and some weren’t; some were likable and some couldn’t be bothered with that. They had styles and perspectives, and each was singular in some way.
The same can’t be said of, say, the bots I encounter on websites when I’m looking for customer service. Aside from their chronic inability to understand what I’m asking, which I assume will get less bad as the technology improves, they’re interchangeable. I forget them as soon as I close the browser in frustration.
Part of the reason that MOOCs didn’t bring the revolution some expected is that education is largely relational. Even when the lecturer on screen is uncommonly eloquent, the relational element is missing. Parasocial relationships aren’t the same as human ones.
Worse, I could foresee digital twins creating workload expectations that effectively rule out actual people. The work would become entirely transactional and impersonal. Faculty would be reduced to the equivalent of the old Scantron machines. Students wouldn’t see their humanity, and lessons wouldn’t stick. We could optimize ourselves into complete system failure.
Albert Finney’s character wasn’t admirable, but he got a basic point right. In this, as in so many things, cheesy ’80s movies will show us the way.