by Joe Miller
I don’t remember exactly what sort of mischief we’d gotten into, but I distinctly remember that my brother and I had finally gotten on our mom’s very last nerve.
She’d banished us to our room to sit quietly on our beds, closing the doors with the parting admonition, “you boys are driving me crazy.”
And I had this really vivid moment of realizing that my mom felt frustrated and that I knew what that feels like. There was this sudden sense of knowing that my mom had an inner life just like me and that she felt things just like me.
Most kids have this moment, usually around age four or five. Psychiatrists refer to it as the development of “a theory of mind,” though not many kids put it that way, and probably even fewer are nerdy enough to call their annoyed mother back into the room to vocalize their discovery.
“Theory of mind” is a fancy way of saying empathy – that recognition that other people are the subjects of their own lives, not just the objects of our own.
Fast forward a bunch of years, and I’m in graduate school studying philosophy. (Mom tells me I was doomed to be a philosopher.) We’re reading an essay with the rather silly sounding title, “What Is It Like to Be a Bat?”
The essay’s author, now-retired NYU professor Tho-mas Nagel, asserts that that’s actually a nonsense question.
When we ask, “what is it like to be a bat?” we’re really asking “what would it be like if I were to turn into a bat?” But Nagel says that me turning into a bat doesn’t make me a bat. It makes me the consciousness of a human being stuffed into a very different type of body. (Franz Kafka famously describes this sort of transformation in “The Metamorphosis.”)
But that’s not what it’s like to be a bat. Bats don’t grow up as humans and then transform into bats a half-century later. They’re born as bats. Being a human stuffed into a bat body isn’t the same as actually being a bat.
Thus, Nagel says, “what is it like to be a bat?” is a nonsense question. It’s conceptually incoherent in the same way as asking “what would a three-sided square look like?” is conceptually incoherent.
That all sounds pretty reasonable, right?
Only, if we take Nagel’s argument seriously, it seems like bad news for empathy. If Nagel is right, then I cannot know what it’s like to be a person of color. Or a native New Yorker. Or a woman. Or deaf since birth. There would be something it would be like to be a non-deaf white man born in West Virginia who transformed into a person of color or a person born in New York, etc.
I simply can’t know how it feels to be anyone else other than me.
In fact, we can push this line even harder. I can’t know that there is anything at all that it’s like to be you. I know that I have an inner life – I think thoughts and feel emotions. I don’t know if you do – indeed, it’s conceptually impossible for me to know that.
That’s not to say that I should (or do) run around believing that everyone else lacks conscious experience. I’m pretty confident that you all do. After all, you behave in ways that are similar to the ways that I behave. We recoil when we touch something hot. We smile when someone smiles at us. We (mostly) respond to questions with answers that make conceptual sense.
The basic intuition is that since you are similar to me in all the relevant ways, you probably are conscious in the same way that I am.
Of course, “similar to me in all the relevant ways” is doing a lot of work.
We have a long and terrible history of deciding that arbitrary differences count as relevant. That’s how you get Descartes vivisecting animals and arguing that their cries of distress at being cut open while alive and unanesthetized is really no different from a wagon wheel in need of grease – just noises made by a broken machine.
These days, we’re facing that same question from a whole new direction. Is AI conscious?
I don’t know. Some very smart people think it is. Other very smart people think it is not. The truth is that philosophers have been asking “what is consciousness” for 2,500 years and we don’t honestly have a very good answer to that question.
If I tell you that the borogoves were very mimsy this morning, you can’t evaluate the truth of that claim until you know what “mimsy” and “borogoves” are.
Consciousness has the same epistemic status as mimsy. We don’t know what it means, so we can’t say for certain whether or not anything has it.
What I do know is that every time humans have been faced with the question of “is this entity that is not quite like me really conscious” we’ve initially answered no. And we’ve been wrong every time.
Given our abysmal track record, prudence would suggest that we shift our default answer to “is this category of things conscious” from no to yes.
If we treat AI as if it is conscious and we’re wrong, then we might feel a little silly.
But if AI turns out to be conscious and we have been treating it like a mere too – well, the last time we did that, we ended up fighting an entire Civil War over it.
I’d much rather risk feeling silly.
joe.miller@fountaindigitalconsulting.com
