In the anime, SAO, Yui, an amnesiac AI construct, is so analogous to a real human girl, that the protagonist mistakes her for being one, and she even forgets initially the fact that she’s a program. With the ability to carry herself in conversation, and traits the very definition of cute, she grows to assume to role of the protagonist’s daughter, and exhibits a personality along with a range of emotions to bolster. You’d be hard pressed to find evidence against Yui’s intelligence inside the anime. In fact, the bonds between the main cast and Yui, is such that it’s almost impossible to think of her as anything less than a fully-functioning human being.
Alan Turing, one of the greatest computer scientist of all time, distilled the question of what it means for a computer to be intelligent to simply, “If a machine acts as intelligently as human being, then it is as intelligent as a human being.”. The eponymous test he proposed to determine the above quality, the Turing test, assesses a computer and it’s programming’s capacity for conversation. With a human participant in an enclosed room, a dialogue between a computer and the participant commences. Pretty simple, but here’s the twist, the participant doesn’t know whether a computer or another person is on the other side. At the end, if he incorrectly guesses that the he just had a conversation with another of his peers, the computer can be said to be intelligent.
Despite the 62 years of research and development since the conceptualization of the test, no program can be said to have a 100% success rate with this. Of course, this doesn’t mean that an “intelligent” computer cannot exist. Cleverbot, an online AI program, never ceases to amaze when confronted with a conversation, even it’s front page declares it an “an AI with some Actual Intelligence”. But while a brief conversation might see it pass the test with flying colors, longer conversations tend to run into some hiccups that give it away. Still, programs like cleverbot, and developments into commercial conversationist programs like Siri, stand as a beacon of hope that one day the test might be cleared, it’s just a matter of future development.
Yui is perhaps the result of these future developments, and then some. Not only is she capable of a more than convincing conversation, but moves, and acts exactly like any other human would. As far as I’m concerned, Yui aces the Asian version of the Turing test. Not only can it be said that Yui acts like a human would, she also seems to be able to think, feel, and love her adoptive parents, Kirito and Asuna. Yui is quite clearly intelligent (at least according to Turing).
But just as the figurative door is about to be closed on the case of Yui’s intelligence, the American philosopher John Searle, puts his foot at the door. “Wait!” He shouts. “Yui isn’t really intelligent!” Being good hosts, we open the door and invite him in for tea.
Still panting, he takes a seat.
Imagine an enclosed room with a man inside. Nobody can see what’s going on in the room, but there’s a slit, the only opening. People start to slip in messages, enquiring if there’s anybody inside. Problem is: the people outside are Chinese, and only write Chinese letters, whilst the man inside the room only speaks and writes in English. Fortunately for him, there’s thousands of cards with Chinese characters on them inside, and a book on how to respond to any message. Unfortunately for him, the book isn’t a English to Chinese dictionary, and can only inform him on what characters to use for a each given message. Consequently, he has no idea what the messages, or his responses really truly mean. For all he knows, he might be part of some sex hotline rig in Hong Kong, while hoping that his messages mean something to the effect of “HELP ME!”.
This is the Chinese room thought experiment John Searle devised to show the disparity between output and semantic understanding. While the Chinese room produces intelligible and varied responses to the age-old question “what are you wearing?”, nothing in the system, not the pieces of paper that comprise the book, not the character cards on the floor, and certainly not the man, truly understand that he’s about to say “Nothing. How ’bout you, sweetheart?” AI programming works similarly- a man in the form of a computer’s processor, takes input from the outside world, uses the book which actually parallels programming, and sends it back out, never truly experiencing anything at all. Yui is simply an infinitely more complicated form of that book.
We close the door, and pour a cup of Earl Grey for Mr. Searle, while preparing to empty our hearts of any sentimentality towards Yui into the fireplace.
“THAT DOESN’T ACCOUNT FOR EVERYTHING!” shouts the reader, who broke the fourth wall just to get here, “If Yui was just programming within a computer, she wouldn’t have been able to have her free will.” he says flailing his arms around, whilst somehow managing to get a portable DVD player, loaded with episode 12, and fast-forwarded to the 14:45 mark in front of us. “HERE, LOOK FOR YOURSELVES!” And like the deranged, Yui fanatic says, the proof is there, explicitly mentioned no less. Yui’s behavior seems totally contrary to what the programmers would’ve wanted her to do, and her ability to ascertain the nature of her existence also proves that she’s self-aware!
Why would a program do something that it was not designed to do? It seems justified to think that it must have free will, if it goes against the programmer’s commands. But there are millions of these occurrences every year in almost every video game, where an NPC takes on the role that it originally wasn’t designed to. I’m not talking about some real life ghost that possesses a character, or how some characters develop consciousness. Nope, I’m talking about the humble, irritating glitch.
Yui’s new found personality isn’t your every day missingno., being something infinitely more adorable, and you might have been beguiled to think that it was something more than a glitch. And with that, let’s return to the man in the Chinese room. Let’s give him a weird enquiry, “Is this the Krusty Krab?” in Chinese. Just as how weird the question was, the response of “No, this is Patrick.” is even weirder. We take a step back and realize that we’ve just illicit a Spongebob reference from a book that’s meant to satisfy the kinkiest of pleasures. But, does this mean that anything in the room has gained an understanding about the immaturity of the people outside? Unfortunately, no.
The Yui fanatic joins us at the fireplace.
As we’re about to clear out the Yui to make more room for the Asuna, a booming voice from beyond the grave calls out, it’s Alan Turing. “Instead of arguing continually over this point, it is usual to have a polite convention that everyone thinks.” (this is an actual quote), and what he’s referring to is the Problem of Other Minds. In a nutshell, if we can’t prove that a being like Yui truly exists, what more can we say about other people, our friends, our enemies, if we don’t have definitive access to their thoughts. The answer is, we, non-philosophers, don’t really care, we go on living our lives assuming that they exist. Perhaps, the glitch that allowed Yui to escape miraculously caused consciousness, perhaps not. We’ll choose the former and hang her posters in our rooms, instead of choosing the latter and hanging ourselves in our rooms.