Intelligence is in the Eye of the Beholder


SelfAwarePatterns has recently made a post titled “Let artificial intelligence evolve? Probably fruitless, possibly dangerous” arguing that if we want to create intelligent machines, we must let them survive in the real world and let them evolve.  I made a comment that may be worth turning into a post.

I said

Intelligence is in the eye of the beholder. “Intelligence”, perhaps, refers to the level of complexity. When a machine is complex enough that we do not understand how it makes decisions to do certain things, we call it “intelligent”. But when we understand how machine’s actions are triggered, the impression of “intelligence” disappears.

For instance, my smartphone may suddenly tell me: “Hey, if you want to be in time for that meeting, you’d better start now and, by the way, avoid that highway – there is an accident near exit 69.” That’s intelligent, right? How did it come up with such a timely and useful message? But, of course, the smartphone “knows” about the time and place of my next meeting from my Google Calendar. It also knows my current location from GPS and can calculate how long it takes to get to the meeting using Google Maps. It also knows about the traffic based on the information from thousands of smartphones on the road aggregated at the Google server. The smartphone does not just “decide” that this message would be useful to me. The smartphone knows nothing of being useful. It is programmed to do things that the designers of Google Now considered useful. So, if we don’t know all these things, the message appears intelligent. But if we do understand how things work, the impression of “intelligence” disappears.

However complex the machine, if it exists, humans (at least, some) must understand how it works. Perhaps, nobody individually, but collectively, there will be a group of experts whose knowledge covers all aspects of the machine. So, perhaps, existing machines will be never considered “intelligent” and the term “intelligent” will always be reserved for some mysterious “next generation”. Of course, nobody has an idea what the next generation of machines will do. So, it’s quite appropriate. On the other hand, we might as well consider that the AI already exists because what I described in my example would certainly blow my mind 20 years ago.

Another thought. “Intelligence” implies purpose. There are very complex natural systems with very complex behavior. But unless they do something that appears useful or purposeful to humans, they are never called “intelligent”. The term “intelligence” seems to be closely related to goal setting and decision making and, therefore, to the question of free will. Before we answer whether machines can be intelligent, we need to answer whether humans are intelligent themselves or are mere automatons. And there is no answer to this question. It’s a matter of philosophical worldview.

 

What does determinism determine?


There is a lot of discussion whether free will is possible in the world where all events are determined by physical laws.   Let’s stop and meditate on the word determined.  What does it mean?  It seems to mean that if we knew the current state of a system with coordinates and momenta of all atoms and molecules, we should be able to know also the state of the system in n seconds from now. There are too many ifs here, don’t you think? First of all, the system must be closed, i.e. confined to itself and isolated from external influences. This implies not only that the system cannot exchange particles with the outside world, but also energy (heat, electromagnetic radiation, or whatever other forms of energy there might be). That’s quite impossible. We might not even be aware of all forms of energy out there considering the “dark matter/dark energy” problem. We also must know the state of each and every atom in this closed system. One atom in the system or coming from outside with unknown state can throw all determinism out of the window.

Just look at the image above.   Can we practically determine the position of every ball after the cue ball hits the pyramid?  The movement of every ball is strictly determined by the Newton’s laws.  What seems to be the problem?  The problem is that we do not know the exact trajectory of the cue ball, where it hits the pyramid.  We do not know the exact alignment of the balls in the pyramid (I am not talking about any quantum principles here).  We do not know how every fiber of the table cloth and every microscopic groove in the table will affect the trajectory of every ball.  We don’t know if all balls are exactly spherical, whether their mass is equally distributed, how many scratches and imperfections they have.  We don’t know if the table is leveled or tilted to one side.  So, OK.  Trajectories of the balls are strictly determined by the Newton’s laws (if we ignore relativistic effects).  That’s what makes this game possible.  But how much is anything “determined”?

And that was a fairly idealistic environment.  Reality usually looks more like this, except the balls are in Brownian motion.

Determinists say that every event is caused by another event.  Let’s stop and meditate on the word caused. What do we mean by that?  Causality is nothing more than another mental construct.  It’s a way to describe relationship between events. In certain cases, when one event follows another event, we say that event 1 caused event 2. But it’s simply a special kind of connection between events.  Which we construct in our mind.  The relationship needs to follow certain rules outlined by Hume (e.g. event 1 must happen prior to event 2, event 2 must always follow event 1, etc. — Hume lists 8 attributes of causality) After spending the last 15 years of my career analyzing “failure root cause” of semiconductor circuits, I realized that that there is no such thing as “the root cause”.  Finding “the root cause” comes down to finding practical and reliable ways of achieving desired results.  Some ways are better than others and some results are more desirable than others.  When a car “accident” happens, we can say that it was caused by distracted driving, slippery road, poor visibility, condition of the car, the tree standing in the way, etc. All of those can be considered as causes of the accident. But there is no magic that determines “THE ROOT CAUSE”.  It’s possible to say that the tree caused the accident because if the tree were not there, the accident would not have happened.  We can even blame evolution that “caused” the tree to exist.  But it does not make any practical sense.  I know the cause of all plane crashes — gravity.  Is this the cause people look for?

car_crash_1

 

We also seem to have freedom to pick and choose the causes of events:

“It all comes,” said Pooh crossly, “of not having front doors big enough.” “It all comes,” said Rabbit sternly, “of eating too much.”

To complicate the matters, an observer is a part of the observed system and inevitably affects the system’s behavior.  Think of turning on the light to see what’s going on in the dark room.  What will you see?  You will see what’s going on in the room right after you turned on the light, not what had been going on in the dark.  In some cases observer effect can be neglected, in some cases, not.  So, the very act of inquiry into natural causes of events adds an element of free will to these events.  Some things seem to be there just because we look.  There is a Russian proverb “If I knew where I fall, I would put some straw there”.    Knowing that an event will happen can change that event.

Now, let’s stop and meditate on the words physical laws. What are they?  Aren’t they also mental constructs allowing us to describe the world and make useful predictions?  Most physical laws operate within a certain idealistic model of the world and ignore “second-order” effects.  What constitutes a “second-order” effect depends on the application.  We can ignore relativistic inaccuracies of Newton’s laws when we play billiards, but when we design a GPS system, we have to take them into account.  When we launch a satellite, we account for the mass of the Earth, perhaps, the Moon, perhaps the Sun.  How about the effects of the gravity pull coming from the billions of distant galaxies?  Probably, not so important.  But second-order effects do exist even if we choose to ignore them.  And we never know when they become too large to ignore.  We also cannot pretend that we know all physical laws.  There are a few that we happen to know.  But I’m certain that there are quite a few of them that we are not even aware of.  Determinism seems to be just another such idealistic model of the world which only makes sense for certain applications, but not others.

Anything left from determinism that has not been reduced to absurdity yet?  Good luck with determinism, Mr. Harris and Mr. Coyne.

See also:

Do we have free will?


Daniel Dennett has recently published a detailed response to Sam Harris’s book “Free Will”.  As one can see from the reviews, the book was received with enthusiasm by many scientists.  E.g. Jerry Coyne, an evolutionary biologist, also chooses the point of view that free will does not exist.  I like how engineer Jacque Fresco explains this idea in the first part of this video (the second part where he mentions how a person may decide to borrow a weapon to protect himself from a wild animal despite being preconditioned by society that stealing is unacceptable seems somewhat inconsistent which is OK considering that, according to himself, he did not choose what to think or what to say).

The popularity of incompatibilism (the notion that deterministic laws of nature are incompatible with free will) among scientists and engineers is not surprising.  It is not a new idea and has been known for many centuries as shown in  this review on Stanford Encyclopedia of Philosophy web site.  All good and well… But…

We still make choices, don’t we?  What do you say, yes or no?  How come?

I tend to agree with Dennett on this issue who espouses compatibilism, the notion that we can consider that free will exists for our practical purposes despite the determinism of laws of nature.  Although Dennett has a few good explanations of his view, I’d like to offer my own.

If we accept that every event is strictly determined by the laws of physics and all circumstances of that event, we must reject also the concept of probability.  There is no likelihood of occurrence of any event because, according to determinism, there is certainty whether and how this event will happen.  Nevertheless, probability is one of the most important concepts in science and everyday decision making.  Why?  I think, it has to do with predictability.  Determinism implies that all events are predictable in principle, if we know all circumstances leading to this event.  However, in practice, we never know all circumstances.  Therefore, we estimate probabilities.  This distinction between in practice and in principle also seems to be the source of confusion about falsifiability of scientific theories.

Let me ask you to name a random integer number between 1 and 10.  Can you do that?  I bet you can.  Now, that you picked a number, it’s not random for you any more — you know what it is.  But for me, the number is still random because I don’t know what number you have picked.  I can still talk about probability of what the number is, whereas for you there is no probability, there is certainty.

Once a lottery has been drawn, there is no longer any uncertainty regarding what the winning numbers are.  The probability of the numbers being what they are is 1.  Probability of the outcome of the lottery only makes sense before the lottery is drawn or if we do not know the results of the drawing.

Imagine a deck of cards face down on the table.  When I take a card from the top of the deck, what is the probability of the card being queen of hearts?  1/52, right?  Now imagine a deck of cards face up on the table with the top card being jack of spades.  What’s the probability of the card being queen of hearts now?  Now it’s zero.  Why?  Because we know it’s not queen of hearts.  But why does it matter whether the deck is face up or face down?  Isn’t the top card what it is regardless of what we think about it?  Why is it reasonable to talk about probability in the first case and unreasonable in the second case?

Perhaps, that’s enough of examples.  In summary, I think, probability, free will, and determinism are  ways to describe the world in a similar way as color is a way to describe the length of electromagnetic waves hitting our eyes (Dennett’s analogy).  In a similar way, imagining that electron is a wave or a particle or forms a cloud is a way to describe certain properties of electron.  But in reality, electron is neither like a wave in the ocean, nor like a cloud in the sky at all.  Thinking that an electron is a particle is incompatible with thinking that an electron is a wave.  Yet, scientists manage to use both models successfully, without contradiction.

It’s important to consider contexts in which concepts are meaningful.  Concepts lose meaning if they are used in a wrong context.  E.g., abbreviation “CD” can mean “compact disc” if we talk about audio records, “certificate of deposit” if we talk about banking, “critical dimension” if we talk about semiconductor processing.  In a similar way, the concept of “free will” is meaningful in most contexts of every day life, but not so in some specific contexts.  “Free will” and “probability” are all mere concepts.  If we argue that they do not exist, we might as well dismiss all other concepts and ideas we may have.

So, although, what I do is determined, I don’t know it, and it gives me the freedom of choice…  Or an illusion of freedom of choice which is as good as freedom of choice for all practical purposes.  There is no difference between free will and the illusion of free will just like there is no difference between real intelligence and fake intelligence.

Related links

 

Negative Capability


Negative Capability, that is, when a man is capable of being in uncertainties, mysteries, doubts, without any irritable reaching after fact and reason Negative capability – Wikipedia, the free encyclopedia Stumbled upon this quote from John Keats.  Interesting concept.  It describes the ability of the individual to perceive, think, and operate beyond any presupposition of […]