Instinct
Knowledge without learning
Recently physicist friend Sofia sent me a link to a story about how a Google engineer named Blake LeMoine had come the conclusion that a google subsystem for language handling called LaMDA was actually intelligent.
https://www.washingtonpost.com/technology/2022/06/11/google-ai-lamda-blake-lemoine/
Google of course denies this. I think that LeMoine is probably right but that both he and Google are philosophically naive about the issue.
Sofia and I discussed this in some detail over a couple of days.
One confusing point is that artificial intelligence (AI) somehow get's morphed into "conscious machine" and then into "can a machine be aware and have subjective experience?" And then we get the tricky move that defines subjective experience as something only living things can have. I don't think tricks like that help at all.
I've found that thinking in terms of cognitive abilities is much more fruitful because it doesn't run into obstacles like that.
By cognitive abilities I mean things like perception and interpretation and awareness and intent.
These are human capabilities for sure, but they are also capabilities that all animals possess to a greater or lesser degree. They aren't unique to humans.
Sofia brought up the point that machines can't have instincts like people. She mentioned that men are sexually attracted to young women as a matter of instinct.
Instinct seems like knowledge directed behavior but can be mindless.
The behavior of the Sphex wasp has been studied by many people. Daniel Dennett uses it as an example of instinctive behavior.
(start of quote from wikipedia)
Some Sphex wasps drop a paralyzed insect near the opening of the nest. Before taking provisions into the nest, the Sphex first inspects the nest, leaving the prey outside. During the inspection, an experimenter can move the prey a few inches away from the opening. When the Sphex emerges from the nest ready to drag in the prey, it finds the prey missing.
The Sphex quickly locates the moved prey, but now its behavioral "program" has been reset. After dragging the prey back to the opening of the nest, once again the Sphex is compelled to inspect the nest, so the prey is again dropped and left outside during another stereotypical inspection of the nest.
This iteration can be repeated several times without the Sphex changing its sequence
(end of quote https://en.wikipedia.org/wiki/Sphex)
The behavior of drop prey, inspect nest, then drag prey in seems to be purposeful, but the wasp has no intent. It's behavior is genetically programmed into it's brain without it having to learn anything.
That is a fabulous fact.
How can something like behavior be determined by genes?
There is a very long story to be told about just how our bodies (brains include) develop from a fertilized egg into an adult human.
For me the obvious conclusion is that our cognitive abilities arise from physical things that do not have those abilities. Our neurons and hormones don't have the capacity to perceive though they enable us to perceive.
When I was a kid it was normal to assume that people were fundamentally different from animals.
We can think and decide while animals could only react instinctively
That view was used to justify all sorts of inhumane and inhuman practices towards animals.
But once the scales fall from the eyes it's seen to be unsupportable.
I've been watching a peregrine nest via webcam for a couple of months now. Also for a while this spring I observed a flock of crows in the grove of trees outside my window.
Birds are pretty amazing but they have very small brains. One might think that the efficient thing would be to pack that brain with automatic responses to all situations so no thought is required.
That requires that the bird have a lot of cognitive abilities like perception and interpretation so that experience is meaningful and can trigger an automatic behavior.
The big problem with a scheme like that is that it doesn't work well in unexpected situations. As I watched the peregrines I could interpret their behavior as instinctive responses to situations. Father bird catches some food and brings it to the nest - instinctive - no thought required.
Mother bird takes the food from father and rips it to shreds and feeds it to chicks - instinctive - no thought required
Chicks can barely move but they can take the bits of food that mother bird offers - instinctive - no thought required.
The chicks grew at a fabulous rate. Since May 5 they've gone from teensy balls of fluff out of an egg to birds as big as their parents.
The chicks now tear their food apart by themselves. I saw one swallow a bit of meat with a long bone attached and it got stuck.
I was like "uhoh now what?" After a couple of minutes mother bird just grabbed the bone and hauled it out of the chick's mouth, pecked the morsel off it and fed that to the chick.
I cheered! And I didn't think that was instinctive behavior. But it was an effective response to an unusual situation.
And my parsimonious idea here is that that response showed a level of thinking and intelligence.
That leads me to the idea that intelligence involves the ability to respond effectively to novel situations.
If we can see intelligence displayed by a peregrine with a very small brain - why not in a machine with similar capabilities of perception and need to interpret novel situations?
Yes I'm thinking of you rovers on Mars and that amazing little helicopter.
But I think what LeMoine had found was that when he presented LaMDA with a novel situation LaMDA came up with a creative and effective response. Isn't that intelligence too?
What do you think?
I present regular philosophy discussions in a virtual reality called Second Life.
I set a topic and people come as avatars and sit around a virtual table to discuss it.
Each week I write a short essay to set the topic.
I show a selection of them here.