It Gets Pretty Weird When You Have Two 'Artificially Intelligent' Chatbots Talk To Each Other
(Turing was also an out-of-the-closet gay man in a time when such a thing was illegal, and when confronted with his "crime," he chose chemical castration over prison time. While the Queen recently granted him a posthumous royal pardon, the modern British government did not exercise the same discretion.)
Turing famously proposed the "Turing test" as a way of gauging a machine's "intelligence." In a Turing test, a human user would interact with a computer via text, exactly as you might use instant messaging to communicate with a friend today. Sometimes there'd be a living, breathing person responding to you, sometimes you're merely communicating with software. If the human user is unable to distinguish between the human and the software respondent, that software is said to have passed the Turing test.
There are loads of chatbots out there today (who else remembers wasting time on AIM by talking to SmarterChild?), and they run the gamut from being totally dumb to dynamic conversationalists. One of the first notable examples is a bot called ELIZA, a project by Joseph Weizenbaum which he developed at MIT between 1964 and 1966. By asking lots of questions and repeating certain key words that came from the user's input, ELIZA could on a certain level simulate speaking to a therapist - ELIZA's most famous "script" was even called DOCTOR. You can chat with ELIZA over the internet right here.
On the other side of the conversational AI spectrum, you have PARRY, developed in 1972 by a psychiatrist named Kenneth Colby. Where ELIZA had its calm, inquisitive nature, PARRY was designed to simulate a schizophrenic. Vint Cerf, one of the fathers of the internet (and incidentally a recipient of the Turing Award) got the two to talk to each other in 1973.
Here are two highlights from that "conversation."
***
PARRY: People get on my nerves sometimes.
ELIZA: I am not sure I understand you fully.
PARRY: You should pay more attention.
ELIZA: Suppose you should pay more attention.
PARRY: You're entitled to your own opinion.
ELIZA: What makes you think I am entitled to my own opinion?
PARRY: It bothers me just to be around people in general.
ELIZA: Please go on.
PARRY: I went to the track at Bay Meadows a while back.
ELIZA: What does that suggest to you?
PARRY: I went to the races.
ELIZA: What else would you like to discuss?
PARRY: Horseracing interests me.
***
PARRY: Do you know anything about bookies?
ELIZA: What makes you think I know anything about bookies?
PARRY: A bookie didn't pay me off once.
ELIZA: What does that suggest to you?
PARRY: I beat up the bookie who did not pay off, and I thought he might use his friends in the underworld to get even with me.
ELIZA: What else would you like to discuss?
PARRY: What do you know about racketeers?
ELIZA: Why do you ask?
PARRY: What experience have you had gambling?
ELIZA: Why does that question interest you?
PARRY: I like horseraces.
ELIZA: What else do you like?
PARRY: I know the mob controls the big rackets.
ELIZA: Who are you really talking about?
PARRY: Have you ever really thought about the MAFIA?
***
It's clumsy, weird, and awkwardly funny. Let's see what happens when we try this with more modern chatbots readily available for use in a web interface. I took the opening line from Cleverbot, fed it to ALICE, then copied and pasted ALICE's response back to Cleverbot. In the screenshot below, "User" is ALICE.
Let's draw out this awkward fun a little longer. I used the same method to get Jabberwacky to talk to Prob.
Artificial intelligence is really, really hard. It's easy to point to systems like Apple's Siri and IBM's Watson as being artificially intelligent, but a lot of smart people have problems with this descriptor. Most notable among them is Douglas Hofstadter.
"Watson is basically a text search algorithm connected to a database just like Google search. It doesn't understand what it's reading. In fact, 'read' is the wrong word," he said in this interview with Popular Mechanics. "It's not reading anything because it's not comprehending anything. Watson is finding text without having a clue as to what the text means. In that sense, there's no intelligence there. It's clever, it's impressive, but it's absolutely vacuous."
I'm not putting these novelty chatbots on the same level as Watson, but this exaggerated example serves the purpose of illustrating Hofstadter's words more clearly. Computer scientists are still plugging away at this, however, even so specifically as to turn the Turing test into a competition. The Loebner Prize has been held every year since 1990, and computer scientists engage in the Turing test exactly as described above with the aim of building a chatbot that can fool participants into thinking it's human. A winner can walk away with $100,000 in prize money, and this all helps plant the seeds for the sci-fi future where we talk to machines like we talk to our friends.
A video from the 2009 competition is embedded below: