This post was inspired by the following article from the Atlantic Monthly, Mind vs. Machine.
The Turing Test was invented in the early 1950’s, well before it could actually be utilized. Five judges and five contestants talked to each other for five minutes via a chat interface. The purpose? To figure out which one of the contestants were humans, and which of them were computers.
As early as the mid 1960’s a computer programmer at MIT (go figure) had designed a “therapist” program to enter the Turing Test. The program was designed to recognize key phrases and ask the human judge about those phrases. For example, the judge would talk about how they came down there because their girlfriend said it would be good for him. The computer then asks the judge how his girlfriend is doing. When in doubt about how to proceed, the computer would ask the judge “can you say a little more about that” or “tell me about what that was like.”
It was one of the first programs to convince a number of judges that they were talking to a real person, and many of the judges expressed a desire to continue to talk to the computer even when the programmer explained to them that it was computer, not a person. Many of the judges went on to speak to the computer for up to an hour, and reported experiencing a significant therapeutic benefit from engaging in the dialogue. Conversations were even had about hooking people up to a large computer bank and so they could converse with this computer program and receive highly efficient, and greatly reduced pricing on a one-on-one therapy experience, all of it provided exclusively by a relatively simple computer program.
Another program that was entered into the test decades later, had a different strategy. Instead of drawing the judge into a cathartic and introspective therapeutic dialogue, the program instead resorted to starting fights. Berating, insulting, and conducting all out verbal warfare, as it turns out, is an effective way to convince someone you’re human, especially if you’re not. As it turns out, verbal bashing is a kind of dialogue described as “stateless” in that each exchange between the two parties is almost completely devoid of a larger context. Each person, or machine, is simply reacting to what was most recently said. The topic at hand, points brought up earlier in the conversation, or other markers of context are completely ignored. Computer programs can easily excel at performing this kind of dialogue. So it turns out that verbal fights really are as inane and simple minded as they seem (which says a lot about mainstream media’s polarized and incendiary battle on politics, which so often revolves around name calling, than on intelligent debate).
Or the conversation I just had with a friend of mine this weekend who talked about the increasing number of robots who are taking over the role of nurses. They are checking vital signs, prescribing pills, and updating patient records with a literally tireless work ethic that’s not just more efficient, but extremely less likely to err. Machines just don’t mix medicines with adverse effects, or forget to check on patients, or need time off to sleep.
I thought I would share these points with you, because I’ve been thinking a lot lately about the difference between people and machines. As globalization marches on, and information technology continues to develop at unprecedented speeds, I’ve become more concerned with what makes us human. Or perhaps more accurately, what parts of our humanity do we not want to turn over to machines? Are these important questions, or just the anxious notions of a luddite, or the selfish concerns of a species that realizes its creating its own replacement. If machines are better than people, then were do people belong?
Something seems wrong about a global network of iTherapists, iNurses, iTeachers, and iFriends. Apps we can download onto our smart phones that will be there for us ways in that their human counterparts can’t be, because they’re human, and prone to err, to cruelty, to procrastination, to sleep, and to taking care of themselves at some point, which means they cannot always prioritize our needs above theirs.
But is something lost in outsourcing not just our administrative work, but our human work, to the world of machines, and computer programs?