One of the most surprising developments in artificial intelligence: it is replacing intellectual labor before manual labor. How can this be possible given that intellectual labor tends to require higher cognitive capacity? How can ChatGPT write in the style of Shakespeare and nearly as well (an almost impossible task) but we don’t have self-driving cars (when most adults are able to drive a car)?
What we didn’t anticipate: the cognitive capacity that separates Shakespeare from the average person is probably .001% of a person’s overall cognitive ability (this is not based on an experiment; I’m just ballparking given that only .1% of our genetics differ from other people, and IQ differences are a very small part of our total genetic differences). Shakespeare and the average person can do basically the same things: see, feel, hear, taste, smell, read, write, understand human emotions, walk, pick things up, etc. In short, Shakespeare and the average person were capable of doing about 99.999% of the same things.
By focusing on just a tiny sliver of intellectual capacity (not learning to process sensory information, how to survive in the world, etc), artificial intelligence has been able to surpass human intellectual ability in many ways. We just had artificial intelligence leap right over 99.999% of cognitive capacity to accomplish it.
But that doesn’t mean AI won’t replace manual labor. It might, ironically, just essentially make humans into robots. Here’s how that might work:
- It’s really hard to teach AI the 99.999% of the cognition necessary to operate in the world in a way that could replace an HVAC technician (getting to the location, ringing the doorbell, maybe calling the homeowner, etc).
- What’s not nearly as hard, though, is for AI to figure out what the problem is once it’s positioned in front of the HVAC system. So, if a technician is wearing a headset that provides video to a cloud AI system, then the AI can evaluate the problem, diagnose it, and verbally speak to the technician to tell them what the problem is and how to fix it.
- If it can be fixed onsite, the AI system doesn’t have to have arms, legs, etc and know how to manipulate them. It can just tell the technician exactly what to do (or, if it’s an easy fix, maybe we’ll just be holding up our phones and AI will tell us directly what to do).
In essence, the human becomes the robot. The robot becomes the brain. In that way, again, AI can skip the 99.999% of cognition that would otherwise be needed to replace a human technician. To put it in even more dystopian language: AI replaces higher cognition, and the human’s own body becomes, essentially, the body of the AI, acting on its directions.
Do I want this future to play out? No. It’s terrifying. The creative destruction will be devastating. But, I don’t think there’s any putting this genie back in the bottle. Best to try to stay ahead of developments as best we can. We were operating from the wrong paradigm (thinking that artificial intelligence had to work up to our highest cognitive abilities). We need to reframe how we think.