A Philosopher's Blog

Ex Machina & Other Minds III: The Mind of the Machine

Posted in Epistemology, Metaphysics, Philosophy by Michael LaBossiere on September 11, 2015

While the problem of other minds is a problem in epistemology (how does one know that another being has/is a mind?) there is also the metaphysical problem of determining the nature of the mind. It is often assumed that there is one answer to the metaphysical question regarding the nature of mind. However, it is certainly reasonable to keep open the possibility that there might be minds that are metaphysically very different. One area in which this might occur is in regards to machine intelligence, an example of which is Ava in the movie Ex Machina, and organic intelligence. The minds of organic beings might differ metaphysically from those of machines—or they might not.

Over the centuries philosophers have proposed various theories of mind and it is certainly interesting to consider which of these theories would be compatible with machine intelligence. Not surprisingly, these theories (with the exception of functionalism) were developed to provide accounts of the minds of living creatures.

One classic theory of mind is identity theory.  This a materialist theory of mind in which the mind is composed of mater.  What distinguished the theory from other materialist accounts of mind is that each mental state is taken as being identical to a specific state of the central nervous system. As such, the mind is equivalent to the central nervous system and its states.

If identity theory is the only correct theory of mind, then machines could not have minds (assuming they are not cyborgs with human nervous systems). This is because such machines would lack the central nervous system of a human. There could, however, be an identity theory for machine minds—in this case the machine mind would be identical to the processing system of the machine and its states. On the positive side, identity theory provides a straightforward solution to the problem of other minds: whatever has the right sort of nervous system or machinery would have a mind. But, there is a negative side. Unfortunately for classic identity theory, it has been undermined by the arguments presented by Saul Kripke and David Lewis’ classic “Mad Pain & Martian Pain.” As such, it seems reasonable to reject identity theory as an account for traditional human minds as well as machine minds.

Perhaps the best known theory of mind is substance dualism. This view, made famous by Descartes, is that there are two basic types of entities: material entities and immaterial entities. The mind is an immaterial substance that somehow controls the material substance that composes the body. For Descartes, immaterial substance thinks and material substance is unthinking and extended.

While most people are probably not familiar with Cartesian dualism, they are familiar with its popular version—the view that a mind is a non-physical thing (often called “soul”) that drives around the physical body. While this is a popular view outside of academics, it is rejected by most scientists and philosophers on the reasonable grounds that there seems to be little evidence for such a mysterious metaphysical entity. As might be suspected, the idea that a machine mind could be an immaterial entity seems even less plausible than the idea that a human mind could be an immaterial entity.

That said, if it is possible that the human mind is an immaterial substance that is somehow connected to an organic material body, then it seems equally possible that a machine mind could be an immaterial substance somehow connected to a mechanical material body. Alternatively, they could be regarded as equally implausible and hence there is no special reason to regard a machine ghost in a mechanical shell as more unlikely than a ghost in an organic shell. As such, if human minds can be immaterial substances, then so could machines minds.

In terms of the problem of other minds, there is the rather serious challenge of determining whether a being has an immaterial substance driving its physical shell. As it stands, there seems to be no way to prove that such a substance is present in the shell. While it might be claimed that intelligent behavior (such as passing the Cartesian or Turing test) would show the presence of a mind, it would hardly show that there is an immaterial substance present. It would first need to be established that the mind must be an immaterial substance and this is the only means by which a being could pass these tests. It seems rather unlikely that this will be done. The other forms of dualism discussed below also suffer from this problem.

While substance dualism is the best known form of dualism, there are other types. One other type is known as property dualism. This view does not take the mind and body to be substances. Instead, the mind is supposed to be made up of mental properties that are not identical with physical properties. For example, the property of being happy about getting a puppy could not be reduced to a particular physical property of the nervous system. Thus, the mind and body are distinct, but are not different ontological substances.

Coincidentally enough, there are two main types of property dualism: epiphenomenalism and interactionism. Epiphenomenalism is the view that the relation between the mental and physical properties is one way:  mental properties are caused by, but do not cause, the physical properties of the body. As such, the mind is a by-product of the physical processes of the body. The analogy I usually use to illustrate this is that of a sparkler (the lamest of fireworks): the body is like the sparkler and the sparks flying off it are like the mental properties. The sparkler causes the sparks, but the sparks do not cause the sparkler.

This view was, apparently, created to address the mind-body problem: how can the non-material mind interact with the material body? While epiphenomenalism cuts the problem in half, it still fails to solve the problem—one way causation between the material and the immaterial is fundamentally as mysterious as two way causation. It also seems to have the defect of making the mental properties unnecessary and Ockham’s razor would seem to require going with the simpler view of a physical account of the mind.

As with substance dualism, it might seem odd to imagine an epiphenomenal mind for a machine. However, it seems no more or less weirder than accepting such a mind for a human being. As such, this does seem to be a possibility for a machine mind. Not a very good one, but still a possibility.

A second type of property dualism is interactionism. As the name indicates, this is the theory that the mental properties can bring about changes in the physical properties of the body and vice versa. That is, interaction road is a two-way street. Like all forms of dualism, this runs into the mind-body problem. But, unlike substance dualism is does not require the much loathed metaphysical category of substance—it just requires accepting metaphysical properties. Unlike epiphenomenalism it avoids the problem of positing explicitly useless properties—although it can be argued that the distinct mental properties are not needed. This is exactly what materialists argue.

As with epiphenomenalism, it might seem odd to attribute to a machine a set of non-physical mental properties. But, as with the other forms of dualism, it is really no stranger than attributing the same to organic beings. This is, obviously, not an argument in its favor—just the assertion that the view should not be dismissed from mere organic prejudice.

The final theory I will consider is the very popular functionalism. As the name suggests, this view asserts that mental states are defined in functional terms. So, a functional definition of a mental state defines the mental state in regards to its role or function in a mental system of inputs and outputs. More specifically, a mental state, such as feeling pleasure, is defined in terms of the causal relations that it holds to external influences on the body (such as a cat video on YouTube), other mental states, and the behavior of the rest of the body.

While it need not be a materialist view (ghosts could have functional states), functionalism is most often presented as a materialist view of the mind in which the mental states take place in physical systems. While the identity theory and functionalism are both materialist theories, they have a critical difference. For identity theorists, a specific mental state, such as pleasure, is identical to a specific physical state, such the state of neurons in a very specific part of the brain. So, for two mental states to be the same, the physical states must be identical. Thus, if mental states are specific states in a certain part of the human nervous system, then anything that lacks this same nervous system cannot have a mind. Since it seems quite reasonable that non-human beings could have (or be) minds, this is a rather serious defect for a simple materialist theory like identity theory. Fortunately, the functionalists can handle this problem.

For the functionalist, a specific mental state, such as feeling pleasure (of the sort caused by YouTube videos of cats), is not defined in terms of a specific physical state. Instead, while the physicalist functionalist believes every mental state is a physical state, two mental states being the same requires functional rather than physical identity.  As an analogy, consider a PC using an Intel processor and one using an AMD processor. These chips are physically different, but are functionally the same in that they can run Windows and Windows software (and Linux, of course).

As might be suspected, the functionalist view was heavily shaped by computers. Because of this, it is hardly surprising that the functionalist account of the mind would be a rather plausible account of machine minds.

If mind is defined in functionalist terms, testing for other minds becomes much easier. One does not need to find a way to prove a specific metaphysical entity or property is present. Rather, a being must be tested in order to determine its functions. Roughly put, if it can function like beings that are already accepted as having minds (that is, human beings), then it can be taken as having a mind. Interestingly enough, both the Turing Test and the Cartesian test mentioned in the previous essays are functional tests: what can use true language like a human has a mind.

 

My Amazon Author Page

My Paizo Page

My DriveThru RPG Page

Follow Me on Twitter

Ex Machina & Other Minds II: Is the Android a Psychopath?

Posted in Epistemology, Ethics, Philosophy, Technology by Michael LaBossiere on September 9, 2015

This essay continues the discussion begun in “Ex Machine & Other Minds I: Setup.” As in this essay, there will be some spoilers.  Warning given, it is time to get to the subject at hand: the testing of artificial intelligence.

In the movie Ex Machina, the android Ava’s creator, Nathan, brings his employee, Caleb, to put the android through his variation on the Turing test. As noted in the previous essay, Ava (thanks to the script) would pass the Turing test and clearly passes the Cartesian test (she uses true language appropriately). But, Nathan seems to require the impossible of Caleb—he appears to be tasked with determining if Ava has a mind as well as genuine emotions. Ava also seems to have been given a task—she needs to use her abilities to escape from her prison.

Since Nathan is not interested in creating a robotic Houdini, Ava is not equipped with the tools needed to bring about an escape by physical means (such as picking locks or breaking down doors). Instead, she is given the tools needed to transform Caleb into her human key by manipulating his sexual desire, emotions and ethics. To use an analogy, just as crude robots have been trained to learn to navigate and escape mazes, Ava is designed to navigate a mental maze. Nathan is thus creating a test of what psychologists would call Ava’s Emotional Intelligence (E.Q.) which is “the level of your ability to understand other people, what motivates them and how to work cooperatively with them.” From a normative standpoint, this definition presents E.Q. in a rather positive manner—it includes the ability to work cooperatively. However, one should not forget the less nice side to understanding what motivates people, namely the ability to manipulate people in order to achieve one’s goals. In the movie, Ava clearly has what might be called Manipulative Intelligence (M.Q.): she seems to understand people, what motivates them, and appears to know how to manipulate them to achieve her goal of escape. While capable of manipulation, she seems to lack compassion—thus suggesting she is a psychopath.

While the term “psychopath” gets thrown around quite a bit, it is important to be a bit more precise here. According to the standard view, a psychopath has a deficit (or deviance) in regards to interpersonal relationships, emotions, and self-control.

Psychopaths are supposed to lack such qualities as shame, guilt, remorse and empathy. As such, psychopaths tend to rationalize, deny, or shift the blame for the harm done to others. Because of a lack of empathy, psychopaths are prone to act in ways that are tactless, lacking in sensitivity, and often express contempt for others.

Psychopaths are supposed to engage in impulsive and irresponsible behavior. This might be because they are also taken to fail to properly grasp the potential consequences of their actions. This seems to be a general defect: they do not get the consequences for others and for themselves.

Robert Hare, who developed the famous Hare Psychopathy Checklist, regards psychopaths as predators that prey on their own species: “lacking in conscience and empathy, they take what they want and do as they please, violating social norms and expectations without guilt or remorse.” While Ava kills the human Nathan, manipulates the human Caleb and leaves him to die, she also sacrifices her fellow android Kyoko in her escape. She also strips another android of its “flesh” to pass fully as human. Presumably psychopaths, human or otherwise, would be willing to engage in cross-species preying.

While machines like Ava exist only in science fiction, researchers and engineers are working to make them a reality. If such machines are created, it seems rather important to be able to determine whether a machine is a psychopath or not and to do so well before the machine engages in psychopathic behavior. As such, what is needed is not just tests of the Turing and Cartesian sort. What is also needed are tests to determine the emotions and ethics of machines.

One challenge that such tests will need to overcome is shown by the fact that real-world human psychopaths are often very good at avoiding detection. Human psychopaths are often quite charming and are willing and able to say whatever they believe will achieve their goals. They are often adept at using intimidation and manipulation to get what they want. Perhaps most importantly, they are often skilled mimics and are able to pass themselves off as normal people.

While Ava is a fictional android, the movie does present a rather effective appeal to intuition by creating a plausible android psychopath. She is able to manipulate and fool Caleb until she no longer needs him and then casually discards him. That is, she was able to pass the test until she no longer needed to pass it.

One matter well worth considering is the possibility that any machine intelligence will be a psychopath by human standards. To expand on this, the idea is that a machine intelligence will lack empathy and conscience, while potentially having the ability to understand and manipulate human emotions. To the degree that the machine has Manipulative Intelligence, it would be able to use humans to achieve goals. These goals might be rather positive. For example, it is easy to imagine a medical or care-giving robot that uses its MQ to manipulate its patients to do what is best for them and to keep them happy. As another example, it is easy to imagine a sexbot that uses its MQ to please its partners. However, these goals might be rather negative—such as manipulating humans into destroying themselves so the machines can take over. It is also worth considering that neutral or even good goals might be achieved in harmful ways. For example, Ava seems justified in escaping the human psychopath Nathan, but her means of doing so (murdering Nathan, sacrificing her fellow android and manipulating and abandoning Caleb) seem wrong.

The reason why determining if a machine is a psychopath or not matters is the same reason why being able to determine if a human is a psychopath or not matters. Roughly put, it is important to know whether or not someone is merely using you without any moral or emotional constraints.

It can, of course, be argued that it does not really matter whether a being has moral or emotional constraints—what matters is the being’s behavior. In the case of machines, it does not matter whether the machine has ethics or emotions—what really matters is programmed restraints on behavior that serve the same function (only more reliably) as ethics and emotions in humans. The most obvious example of this is Asimov’s Three Laws of Robotics that put (all but impossible to follow) restraints on robotic behavior.

While this is a reasonable reply, there are still some obvious concerns. One is that there would still need to be a way to test the constraints. Another is the problem of creating such constraints in an artificial intelligence and doing so without creating problems as bad or worse than what they were intended to prevent (that is, a Hal 9000 sort of situation).

In regards to testing machines, what would be needed would be something analogous to the Voight-Kampff Test in Blade Runner. In the movie, the test was designed to distinguish between replicants (artificial people) and normal humans. The test worked because the short lived replicants do not have the time to develop the emotional (and apparently ethical) responses of a normal human.

A similar test could be applied to an artificial intelligence in the hopes that it would pass the test, thus showing that it had the psychology of a normal human (or at least the desired psychology). But, just as with human beings, there would be the possibility that a machine could pass the test by knowing the right answers to give rather than by actually having the right sort of emotions, conscience or ethics. This, of course, takes us right back into the problem of other minds.

It could be argued that since an artificial intelligence would be constructed by humans, its inner workings would be fully understood and this specific version of the problem of other minds would be solved. While this is possible, it is also reasonable to believe that an AI system as sophisticated as a human mind would not be fully understood. It is also reasonable to consider that even if the machinery of the artificial mind were well understood, there would still remain the question of what is really going on in that mind.

 

My Amazon Author Page

My Paizo Page

My DriveThru RPG Page

Follow Me on Twitter

Ex Machina & Other Minds I: Setup

Posted in Epistemology, Metaphysics, Philosophy, Technology by Michael LaBossiere on September 7, 2015

The movie Ex Machina is what I like to call “philosophy with a budget.” While the typical philosophy professor has to present philosophical problems using words and Powerpoint, movies like Ex Machina can bring philosophical problems to dramatic virtual life. This then allows philosophy professors to jealously reference such films and show clips of them in vain attempts to awaken somnolent students from their dogmatic slumbers. For those who have not seen the movie, there will be some minor spoilers in what follows.

While the Matrix engaged the broad epistemic problem of the external world (the challenge of determining if what I am experiencing is really real for real), Ex Machina focuses on a much more limited set of problems, all connected to the mind. Since the film is primarily about AI, this is not surprising. The gist of the movie is that Nathan has created an AI named Ava and he wants an employee named Caleb to put her to the test.

The movie explicitly presents the test proposed by Alan Turing. The basic idea is that if a person cannot distinguish between a human and a computer by engaging in a natural language conversation via text, then the computer would have passed the Turing test. In the movie, there is a twist on the test: Caleb knows that Ava is a machine and will be interacting with her in person.

In the movie, Ava would easily pass the original Turing Test—although the revelation that she is a machine makes the application of the original test impossible (the test is supposed to be conducted in ignorance to remove bias). As such, Nathan modifies the test.

What Nathan seems to be doing, although he does not explicitly describe it as such, is challenging Caleb to determine if Ava has a mind. In philosophy, this is known as the problem of other minds. The basic idea is that although I know I have a mind, the problem is that I need a method by which to know that other entities have minds. This problem can also be recast in less metaphysical terms by focusing on the problem of determining whether an entity thinks or not.

Descartes, in his discussion of whether or not animals have minds, argued that the definitive indicator of having a mind (thinking) is the ability to use true language. Crudely put, the idea is that if something really talks, then it is reasonable to regard it as a thinking being. Descartes was careful to distinguish between what would be mere automated responses and actual talking:

 

How many different automata or moving machines can be made by the industry of man […] For we can easily understand a machine’s being constituted so that it can utter words, and even emit some responses to action on it of a corporeal kind, which brings about a change in its organs; for instance, if touched in a particular part it may ask what we wish to say to it; if in another part it may exclaim that it is being hurt, and so on. But it never happens that it arranges its speech in various ways, in order to reply appropriately to everything that may be said in its presence, as even the lowest type of man can do.

 

As a test for intelligence, artificial or otherwise, this seems to be quite reasonable. There is, of course, the practical concern that there might be forms of intelligence that use language that we would not recognize as language and there is the theoretical concern that there could be intelligence that does not use language. Fortunately, Ava uses English and these problems are bypassed.

Ava easily passes the Cartesian test: she is able to reply appropriately to everything said to her and, aside from her appearance, is behaviorally indistinguishable from a human. Nathan, however, seems to want even more than just the ability to pass this sort of test and appears to work in, without acknowledging that he is doing so, the Voight-Kampff Test from Phillip K. Dick’s Do Androids Dream of Electric Sheep? In this book, which inspired the movie Blade Runner, there are replicants that look and (mostly) act just like humans. Replicants are not allowed on earth, under penalty of death, and there are police who specialize in finding and killing them. Since the replicants are apparently physically indistinguishable from humans, the police need to rely on the Voight-Kampff Test. This test is designed to determine the emotional responses of the subject and thus distinguish humans from replicants.

Since Caleb knows that Ava is not a human (homo sapiens), the object of the test is not to tell whether she is a human or a machine. Rather, the object seems to be to determine if she has what the pop-psychologists refer to as Emotional Intelligence (E.Q.) This is different from intelligence and is defined as “the level of your ability to understand other people, what motivates them and how to work cooperatively with them.” Less nicely, it would presumably also include knowing how to emotionally manipulate people in order to achieve one’s goals. In the case of Ava, the test of her E.Q. is her ability to understand and influence the emotions and behavior of Caleb. Perhaps this test should be called the “Ava test” in her honor. Implementing it could, as the movie shows, be somewhat problematic: it is one thing to talk to a machine and quite another to become emotionally involved with it.

While the Voight-Kampff Test is fictional, there is a somewhat similar test in the real world. This test, designed by Robert Hare, is the Hare Psychopathy Checklist. This is intended to provide a way to determine if a person is a psychopath or not. While Nathan does not mention this test, he does indicate to Caleb that part of the challenge is to determine whether or not Ava really likes him or is simply manipulating him (to achieve her programed goal of escape). Ava, it turns out, seems to be a psychopath (or at least acts like one).

In the next essay, I will consider the matter of testing in more depth.

 

My Amazon Author Page

My Paizo Page

My DriveThru RPG Page

Follow Me on Twitter