[OTL] The Turing Test in BSG
Before I discuss the relation of Turing’s Ideas with Battlestar Galactica I will first provide some background.
Alan Turing was a pretty intelligent guy to say the least. Known as the father of theoretical computer science and artificial intelligence, he was one of the most influential minds of the twentieth century. He developed some of the first computers and computer software and is even credited with cracking the Enigma code, a secret language that the Nazis used when transmitting military messages during WWII. Aside from his work in computer science and cryptanalysis, he was also very well known for his work in the field of Artificial Intelligence Philosophy. His paper, “Computing Machinery and Intelligence”, focuses on addressing the question, “Can Machines Think?”, a question that becomes more and more relevant as computers become faster and more versatile. For the crazed Battlestar Galactica addict, this question is probably even more enticing, since the show presents us with humanoids who were made by machines with digital brains. Before I get into this though, I will try to first give a quick rundown on some of the most important themes of in Turing’s paper.
Right in the beginning of the paper, Turing pretty much throws away the original question of “Can machines think?” since it is a kind of vague question. Instead of this question he proposes a different kind of question by first constructing a thought experiment. This experiment involves a form of the imitation game known as the Turing Test. In this game, both a machine (digital computer) and a human are asked a series of questions by an interrogator, who tries to guess which responses are from the human and which responses are from the computer. The machine’s job is to try to get the interrogator to guess wrong. Thus, in order to be successful, the machine’s cognition needs to appear to function in a nearly indistinguishable manner to that of a human being. Turing, being very pro-machine, concludes that machines have the capability to succeed at the Turing test in the future. According to him, the only thing limiting them is storage capacity. Turing proceeds to go about proving his thesis by disproving some opposing arguments. From each of his dismissals it appears that the overall theme is that we should judge machines with the same standards that we do for humans.
Turing’s response to objection number five exhibits his “fair play for machines” argument quite well. Turing frames objection number five like this, “I grant you that you can make machines do all the things you have mentioned but you will never be able to make one to do X”. X could be something like, “a machine can never exhibit the same diverse set of behaviors that a human can!”. Whoever believes this, Turing argues, is using inductive reasoning. To inductively reason is to take a set of specific facts and from those draw a general conclusion about the topic at hand. For example, if I was a Cylon who (that??) liked to inductively reason and had just encountered Gaius Baltar as the first human I had ever met, then I would conclude that all humans were creepy, selfish, scientific geniuses, and lust-driven creatures. This is definitely false. We are not all like Gaius Baltar! Anyway, if we go back to the original claim that a machine is incapable of showing diverse set of behaviors, Turing would claim that this is just an assumption based on our experience with current machines. In fact, with the right programming and storage capacity a machine could certainly act in ways equivalent to those of a human.
Perhaps better example of false inductive reasoning in regard to machines is the claim that “machines can never make mistakes! Therefore, they are not human!”. According to Turing, an ideal machine can certainly make mistakes. A machine could easily fall into the type of mistakes that Turing calls, “Errors of conclusion”, which are just errors due to an incorrect method of thinking. These are the types of errors humans make every day. Conclusions by inductive reasoning can turn out to be errors of conclusion for example. There is no reason a machine can’t make this exact type of mistake with the right program, and thus make the same type of mistakes a human might make when asked questions during the Turing test.
Although I only went through one of the objections that Turing attempts to rebut I think it demonstrates his goal pretty clearly, in that we should avoid unfounded assumptions and judge Machines just like we would judge ourselves. Fair play for machines!
Now let’s get into battlestar galactica. The humans of battlestar galactica definitely want to believe that they operate on some sort of a higher level of thinking than machines do. One scene that captures this assumption is in the episode “Flesh and Bone” where Starbuck is interrogating Leoben. During part of this confrontation both try to prove that their “species” is intellectually superior to the other. During the interrogation, Starbuck taunts the Cylon, telling him if that he could somehow turn off his ability to feel pain it would degrade him from the status of personhood to that of a machine. She states, “if turn off the pain now and you will feel better, but that makes you a machine, not a person.” Presumably, Starbuck hopes that Leoben will equate the status of “machine” with operating at an inferior level to humans. Turing would object to the assumption that operating like a “machine” disqualifies one from acting human because according to him, machines can certainly emulate the thought process of humans. He would probably accuse Starbuck of using prior experience with computers to inductively reason that digital machines are not capable of operating with a human thought process.
Another thing to note: The humanoid Cylons in Battlestar Galactica are definitely exempt from the Turing test since they are not digital and have actual brains. Turing specifically states that the Turing test is supposed to test digital machines and that any biologically cloned human would probably pass the Turing test easily. Despite this, it is a common trend for a lot of the Galactica members (like Starbuck) to label the Cylons as just “machines” to prove that they are inferior and would fail the Turing test. Turing would say that this label does nothing to prove that Cylons would fail, since, as he argues, digital machines could pass the Turing tests.