Testing Turing

Turing Tests

 

 

The classic Turing test involves a robot that passes for a conscious human. The examiner spends time with the robot, asking questions, interacting, and the question is whether it presents a convincing appearance of intelligence and consciousness. It is like an audition for playing the part of a normal human being. Structurally, however, the Turing test exemplifies something more general, and it is instructive to spell out what this is.

Thus consider the Turing* test: can we construct a virtual world that passes for a real world? An engineer is making a machine that will feed inputs into the brain and produce an impression of a world of ordinary material objects; the question is whether this virtual world can convince someone that it is real. The subject can experiment on this virtual world, moving around, varying the angles, using different senses, and if after some suitable time cannot distinguish the virtual from the real, we can declare that the machine passes the Turing* test. It can produce a convincing simulacrum of a real world—as a robotics engineer might produce a convincing simulacrum of a conscious intelligence.

We could also envisage a Turing** test that concerns producing artificial plant life: can we make an object that resembles a naturally occurring plant enough to convince someone that it is really a biological plant? And we can have subdivisions of such questions: can we artificially simulate a virus, a bat, a cactus, or an octopus? The question is not specific to robots and minds at all: it is about the power to mimic naturally occurring objects by artificial contrivance. Can we make an artificial F, for arbitrary F?

Here is an interesting question of this general type—call it the super-Turing test: can we create a virtual world that contains robots that pass the classic Turing test? That is, we first have to create a virtual world of bodies, as in the Turing* test, and then we have to ensure that those virtual bodies behave in ways that perfectly mimic human bodies—so that they will pass the Turing test. Thus virtual robots may pass the super-Turing test, and hence be declared conscious intelligent beings.

Suppose they do pass that test: are they then really conscious? But how can a merely virtual being be conscious? Are people in your dreams conscious? Passing the Turing test is not logically sufficient to qualify as conscious, because passing the super-Turing test is not sufficient. Passing the test is enough to convince someone that there is a real thing here of the type in question, if they don’t know the actual nature of thing; but that is a question about evidence and belief, not about what is metaphysically possible. Anything can pass a Turing-type test for being an F but still not be an F.

 

 

Ex Machina

I went to see the new film Ex Machina, about whether a female robot will pass the Turing Test. It is never quite clear whether she is conscious in a human kind of way, even when she is flirting with her examiner. But the matter is put beyond doubt when she turns on her maker and stabs him to death, and then leaves the man she flirted with to die of starvation. It is the existence of evil that proves she is conscious–that she is human. Also, the film contains a nice exposition of Mary in her black and white room. All very philosophical.

Stalin’s Fallacy

Stalin famously said, “Death solves all problems–no man, no problem”. He meant that murdering people solves the problem their existence poses–hence his murder of millions of Russians who threatened to be a “problem”. I suspect many people in power have thought along similar lines–and the saying has a certain cogency. But it is important to see that it is a fallacy, a shallow fallacy. Consider Socrates: he is still a problem for those who oppose free thought and the questioning of authority. The dead can still be symbols of what the oppressor wishes to destroy–potent symbols. Alan Turing is a potent symbol, as is Oscar Wilde (actual murder may not be necessary to get someone out of the way). The dead can still cause problems, by their memory. There are many obvious historical examples. Stalin was wrong.

Philosophy

Why does philosophy even exist? Is the world an inherently philosophical place? That seems unlikely. Is it our concepts that generate philosophical problems? But why should we have concepts that do that? Does it show there is something deeply wrong with our concepts? And why are we so confused, if conceptual confusions are the problem? We don’t seem all that confused about other things. One the main problems of philosophy is its own existence.

Philosophical Fun for the Whole Family

The Simulation Game

 

 

The following document recently fell into my hands:

 

Report to the Commissioner of Games: We recently met to discuss, plan and implement a new game, to be called the Simulation Game (hereafter SG). For this purpose we have created a small group (about 6 billion) of individual centers of consciousness, each with finite and quite restricted intelligence. These are to be the pieces in the game and they are currently stored in warehouse 7,000,042, suitably hooked up to the simulation machine SM 5000. The system is now fully operational, with each individual experiencing a fully simulated world. Our technicians have verified that there are no glitches.

Each individual believes that he or she is living in a world that really exists. The point of SG is to provide clues to the pieces that this is not so and see when they realize they are in a simulation. We considered inserting some obvious clues into their stream of experience, such as sky writing that says “This all a simulation—you are being fooled”, but that was deemed a bit too obvious, even taking into account the limited intelligence of the pieces. To make the game more interesting, and to net the greatest gambling revenues, we decided to make the clues subtler, though of course any of our species would recognize them immediately. We have therefore arranged it so that the world they experience is incoherent and unintelligible—quite literally impossible. This is not so clear on the surface, but in the game it is meant to be gradually revealed, as they apply their limited intelligence to the appearances.

The bets are on who will get there first, if anyone. Without going into unnecessary detail, we have built into the simulation a few telltale incoherencies—such as the idea that consciousness depends upon the brain, some logical paradoxes, and the measurement problem (etc) surrounding quantum physics. In SG the pieces are allowed to discover and reflect upon their “world” and to ask themselves whether it really makes any sense. Once they realize it doesn’t, the question is when they will hit on the correct explanation of their predicament: that they are pieces in a simulation game. So far the vast majority are clueless but a few have begun to suspect that all is not well—they are starting to feel that they live in an impossible world (or “world”). At later stages of the game the point of interest will be whether they can persuade others of the truth.

SG promises to be quite fun and completely harmless (unlike that game Galaxy Busters dreamt up recently by some irresponsible gamers). We ask merely that you, as commissioner, list the new game in your records and grant us the appropriate patent. Thanking you for your attention, we are the Grand Society for Responsible Gaming, Section 345, Plasma System 68,000,333.”

 

What should we make of this peculiar document? First, if it is genuine, then these super-gamers are by no means infallible, since the document gives the game away completely. But perhaps they are just being clever, since the existence of such a document by means proves the truth of what it contains. So let us put that aside. The story seems perfectly intelligible: it is logically possible to create a simulated world that contains hidden incoherencies—as with many works of fiction or even dreams. Thus there can be internal evidence that a narrative is a form of fiction not fact. A simulation might undermine itself in this way, either by design or through incompetence. We might think that the authors’ breezy assumption of incoherence in the simulated world they have created is questionable. Admittedly, the puzzles of mind and body, of the logical paradoxes, and of the quantum realm are serious and hitherto resistant to intelligible solution, but maybe this is a just a matter of temporary perplexity, or perhaps of permanent cognitive limitations on our part. Why should we think that a world in which these problems arise is impossible? Well, that depends on how seriously unintelligible you think the world is—whether such a world would be genuinely impossible. Of course, a world cannot be impossible if it exists to be lived in; but it may be that our “world” is unintelligible simply because it does not exist. That is, there are no bodies and brains and objects of the kind that we suppose, governed by the laws we think we have discovered. It is an impossible fictional world—a perfectly intelligible notion. This is the view of the designers of the simulation game: such a world is literally impossible and the question is whether we will come to realize this and draw the obvious (to them) conclusion, namely that we are pieces in a game of simulation. The form of the argument is straightforward: unintelligible worlds cannot exist; our (apparent) world is unintelligible; therefore our world does not exist. Given that we experience an apparent world, the best explanation is that we are living in a simulation contrived by superior aliens.

Question: does this story raise the probability that we are subjects in a simulation game? Suppose that we are: do we now have clues that this is our situation, if only we interpret them rightly? If our true situation is revealed tomorrow, will we be right to say, “Yes, it was clear all along, if only we had heeded the signs”. I just wish the document had contained information about what would happen if we arrive at the correct view of our condition. What happens to us at the end of the game? Will it be a case of “game-over” with all the pieces disposed of, or will we be put out to pasture in some undreamt of paradise?

 

 

Blind Review

The obsession with blind review of journal articles is peculiar. Why not blind examining or blind book reviewing or blind job selection? Anyone entrusted with these jobs is expected to evaluate in an unbiased manner, and if they can’t they should not be so trusted. I think all information about a candidate can be useful, including their identity. If I am reviewing an article by an unknown author I take particular care to ensure that they get a fair hearing. It can also be useful to know if an article is by someone established. In addition, blind review is often a sham, because information leaks out. Insistence on blind review effectively says to the reviewer, “You can’t be trusted to act in a fair manner”.