Key points
- AI can make us more capable while asking us to think less.
- The real shift may begin when thinking simply isn’t necessary anymore.
- What happens when the easier choice is to let the machine do it?
Three years ago, I wrote a post asking whether artificial intelligence might eventually place human cognition on the technological chopping block. At that time, AI was beginning to reveal what it could do, and the concern seemed fairly straightforward. Innovation has always created obsolescence. Machines replaced physical labor, calculators took over much of routine arithmetic, and digital technologies displaced entire industries. But now the domain of innovation is less about that old stereo or calculator but human thought itself.
OK, cognition hasn’t become obsolete per se. In fact, AI has made us astonishingly more capable. Scientists can explore hypotheses faster, physicians can interrogate complex medical information, and you and I can analyze, write, code, or create at levels that previously required specialized (and expensive) expertise. It’s truly remarkable and I don’t think we should diminish it simply because we’re concerned about some of the consequences
I’ve called one expression of this Iterative Intelligence—the back-and-forth between human and machine in which an idea gets questioned, reframed, and sometimes pushed somewhere neither would have taken it alone. And while this is new, it also harkens back to the venerable Socratic dialogue. Used in this context, the technology doesn’t necessarily reduce thinking. Sometimes it creates more of it. So I’m increasingly uncomfortable with the simple question of whether AI is making us smarter or dumber. Something more complicated may be happening.
Obsolescence Begins With Necessity
We tend to think of obsolescence as an endpoint. Something works until a better technology arrives and eventually the old thing gets thrown away. Film cameras, fax machines, and typewriters come to mind. But before any of them disappeared, something more consequential happens and this is important to recognize. We simply stopped needing them. And perhaps that’s a better way to think about cognition in the context of AI
AI doesn’t have to destroy my ability to write. It only has to make writing something myself less necessary. My memory doesn’t have to deteriorate if remembering becomes less important. And my ability to construct an argument doesn’t have to disappear if a reasonable argument can be generated before I’ve even worked through it
This leads me to a concept I think of as Subtractive Innovation. That’s wheninnovation increases the capability of the human-machine system while reducing the necessity for the human to exercise the capability being augmented. And I think the word necessity is critical here as I’m not arguing that using AI inevitably makes us cognitively weaker. Human beings have been offloading cognitive tasks onto tools for thousands of years, and many of those tools have made us smarter and more capable. AI may do the same on a scale we have never seen before.
So, here’s what concerns me. Human abilities tend to have a “use-it-or-lose-it” quality. When we don’t need to do something, we generally do it less. And if we do it less, I think it’s reasonable to wonder what happens to the underlying capability over time. I don’t think we know the answer yet, but the pathway is certainly worth paying attention to
When Thinking Becomes Optional
A blank page once required someone to write the first sentence. An argument had to be constructed, and even asking a good question usually required enough understanding to know what you didn’t know
So, here’s the point: Increasingly, these things are optional
AI can start the page, finish it, and summarize it. And here’s where it gets interesting: Every one of those choices can be completely rational. Why spend 20 minutes doing something that can be accomplished in 20 seconds, particularly when the result may be just as good or even better?
This is where I’d suggest that Subtractive Innovation becomes different from simple cognitive offloading. AI isn’t taking our ability to think away from us. It’s removing, piece by piece and thought by thought, some of the necessity to think for ourselves. Cognitive surrender, if it happens, probably won’t come from some dramatic decision to turn our minds over to machines. It may emerge almost unnoticed through thousands of sensible decisions to avoid effort that technology has made unnecessary.
Artificial Intelligence Essential Reads
The New Wire(d) Mother
Could You Fall in Love With an AI Chatbot?
The Choice Inside the Machine
Iterative Intelligence offers another path. Instead of asking AI to eliminate the cognitive act, we can use it to extend that act. We can challenge its answer, change the premise, find the contradiction, or follow an association we hadn’t considered. Two people can use exactly the same AI and arrive at equally impressive outputs while having very different cognitive experiences. One may have used the machine primarily to complete the task, while the other used it to change the thought. Curiously, this path isn’t the shortest distance from question to answer, it’s the longer and bumpy experiential path of discovery and learning.
This brings me to the idea of abundance, another word that has become almost inseparable from the promise of AI. I’ve written before about the strange dynamics of cognitive abundance, and I think Subtractive Innovation adds another dimension to it. We tend to assume the math moves in one direction—more intelligence available to us should mean more intelligence for us
I’m not sure the equation is that simple
The abundance of machine capability can make the human-machine system dramatically more powerful while simultaneously making some human capabilities less necessary. Both things can be true at once. We can have more intelligence available to us and fewer reasons to exercise our own
A technological lifetime ago, about three years, I wondered if cognition itself might eventually become obsolete. Today, I think obsolescence may be the wrong place to begin looking. The earlier signal is necessity. A capability can remain perfectly intact while becoming progressively less necessary, and perhaps that’s where the real cognitive consequences begin
AI may indeed usher in an age of abundance. But the math may not move in only one direction. And more at the level of the system doesn’t necessarily mean more at the level of the human
And that leaves me with a slightly different question. What happens to humans when thinking becomes optional?
