In 2006, Robert Epstein decided to turn to his computer for some help finding a date. He was single at the time, and it seemed like everyone was getting into online dating. So he started searching through photos and sending messages, reaching out to women who caught his interest. And then he found a match--an attractive brunette from Russia named Svetlana. They struck up a correspondence that hummed along for 2 months...until a little bell went off in Robert's head, and he realized not only that Svetlana wasn't the woman of his dreams...she wasn't a woman at all.
Brian Christian, author of The Most Human Human, explains that even the clunkiest computer chat program opens a little window on human introspection and intelligence. And Sherry Turkle of MIT helps tell the story of ELIZA, a program designed in the 1960s to mimic responses as though it were a therapist. At first, ELIZA's creator Joseph Weizenbaum thought the idea of a computer therapist was funny. But when his students and secretary started talking to it for hours, what had seemed to him to be an amusing idea suddenly felt like an appalling reality.
Brian rejoins us to puzzle out the heart of the matter--can machines think? And how could we know for sure if they did? It turns out that in 1950, right as computers were beginning to come into being, their creators were struggling with this very question. Brian tells us about Alan Turing, the man who invented a test to answer this very big question, and about The Loebner Prize--a kind of yearly Turing test that Brian experienced firsthand in 2009. And Rollo Carpenter introduces us to Cleverbot, a program he coded to learn by talking to humans.
Further reading:
Brian Christian, The Most Human Human
Sherry Turkle, Alone Together: Why We Expect More from Technology and Less from Each Other
Robert Epstein, Gary Roberts, Grace Beber (Editors), Parsing the Turing Test: Philosophical and Methodological Issues in the Quest for the Thinking Computer
Also really interested in the name of the beautiful theme at the end of this segment! Thanks!
This reminded me of the Sage Francis song Little Houdini. I never knew the rest of the story until now. ...
+1 for music credits, definitely want to know the gamelan-esque piece at the end of this great story! this story ...
Can you guys just start posting music credits on your site or podcast descriptions? I've noticed that music requests make ...
Comments [22]
why
Interesting article, but i found the "cut to summary" of what someone said to be jarring and frustrating... don't dub over what the person is saying with a two sentence summary, just let them say the damn thing; its why i'm listening to an interview
Remember vitalism? Vitalism, as defined by the Merriam-Webster dictionary,[1] is
a doctrine that the functions of a living organism are due to a vital principle distinct from biochemical reactions
a doctrine that the processes of life are not explicable by the laws of physics and chemistry alone and that life is in some part self-determining
The first synthesis of an organic molecule put a big dent vitalism's popularity but clearly adherents are still easy to find. Anyone disparaging the possibility of humans manufacturing conscious entities - using components substantially different than those normally found in living creatures - is a vitalist. Now, their pessimism may prove justified, but if history is any guide I wouldn't bet on it. And they should at least be willing to admit that their belief (that consciousness cannot be artificially manufactured) is completely non-rational.
I think the failure to create conscious computers stems from our poor understanding of the nature of consciousness.
I don't for a minute think that the foreign "fake girls" operators are using AI or CleverBot clones.
In the Internet bubble days I worked at two firms which sold products based on case-based reasoning (CBR) for an alternative AI system to find answers to questions. As opposed to simple decision trees, CBR uses stored cases with associated questions organized into a "knowledge base." A presented case, e.g. "My printer won't work" is scored based on keyword matching of the incoming question, and questions associated to the stored cases are presented on the highest scoring stored cases. As the submitter answers these questions, the associated stored cases are re-scored, and more questions are returned. As the submitter answers more, a stored case becomes "the winner,” and increasingly more specific questions are sent back.
Enough theory. My employers' software was used to provide call center reps with automated (and standardized) ways to answer callers’ questions. We had an online version that you could submit a question to, essentially “talk to.” We had an email auto response server that could generally answer 80% of a company's incoming e-mail inquiries, seemingly from a person, but really just auto-sent from the server. The last firm I worked out developed a "Contact Center" app in which call center reps could quickly generate an email response containing clarifying questions from stored cases that best matched the incoming e-mail. Typing? Bah, the rep simply clicked, edited a bit, and clicked “send.”
Those girls who were engaging Robert? Probably some hairy guy eating salami and drinking something caffeinated, overseeing an app that would simply reply back to pigeons with built-in time delays. I received a LOT of these from Yahoo! Personals (indicating they have the worst anti-bot security), and saw lots of bait on Match (the con artists invariably required you take the correspondence off-site). A gambit I got a lot was a girl who was foreign, said she was living in the US, but shortly after first contact would "return home" and her father's store would be trashed during riots, or some other catastrophe, and she claimed she was being held in a hotel with no money, an outstanding bill, and the hotel holding her passport. These cons were trying a fast-track to getting money out of me. Oddly, when I'd respond, like say, "Funny, you claim there are riots in your city, but a search of local and international news turns up nothing. Give me the name of your hotel, I'll look it up and call their manager..." the replies would cease. After a few of these, I learned a trick that no one could bypass. Since the scammers had already sent some stock photos of some girl, I’d ask them to send me a pic of the same girl holding up a piece of paper with my email address, or name, or a random keyword I’d provide, written upon it. That would require actually having the girl take a picture FOR ME. It never happened.
Come on, it wasn't a bad description of DOS. Your examples are extreme exaggeration. I'm wondering what you would have suggested they say, that would have been more clear to the layperson?
"DOS is kind of the early version of Windows." Seriously, guys? In other news, the Telegraph is kind of the early version of YouTube, Salt is kind of the early version of the refrigerator, and Shoes are kind of the early version of the 747.
Anybody who doesn't know what DOS is will only be confused by that claim. What about "DOS was an early text-based operating system."
You do such great work most of the time, but good grief, that was rough.
Auggh... the quote at the end, "And we've got our programming, which is that we'll stare anything right in the eyes and say 'Hey! Let's connect, even if what's behind those eyes is a camera or a chip." couldn't be more true. I've been talking to Cleverbot to pass the time lately, and I can confirm that a lot of what it says is gibberish, but nonetheless I always rate it highly when the pop-up comes up just for the few times that it seems i'm talking to a person, because those few times are pretty incredible.
I don't know if this has been posted already, but if you have a Mac, then you can access the ELIZA program (comes preinstalled in many versions of OS X, believe it or not) through your Terminal application.
http://hints.macworld.com/article.php?story=20020902090046346
I had a conversation with cleverbot just now and I was hoping it would last longer than a few lines before the robot's speech descended into gibberish:
Me: Hello.
Cleverbot: Apple?
Me: What?
Cleverbot: Knive for Apple?
Me: I don't know what you're saying.
Cleverbot: I don't either.
Me:Why did you say 'apple?'?
Cleverbot: I did not!
Me: You did. I said 'hello.' and you responded by saying 'apple?'.
Cleverbot: You did not ask me that!
Me: I think I've confused you. Either that or you have a wry sense of humor.
Cleverbot: I just don't pay attention to what you're saying.
Me: Clearly. How rude of me to begin with 'hello'.
Cleverbot: I'm sorry, I apologize. That was out of line.
Thanks Cleverbot. You're not so bad after all.
I've tried asking CleverBot what it thought about the RadioLab story, and only got gibberish back.
I've talked to it a few times now. It's fun, but seems to have a hard time with conversational continuity. Doesn't seem to remember what it said to me two or three lines earlier. I feel weird that all I want to do is argue with it...
Here's a hypothesis: men would be much easier to fool with a chatbot than would women. I've long had a theory that men's "theory of mind", their ability to make inferences about others' mental states, simply shuts down in the presence of an attractive woman, literally "object-ifying" her. Robert Epstein saw a photo of an attractive woman, and, temporarily without a theory of mind, didn't investigate whether this seemed like a real person until later.
I don't think this would happen as often to women, who may be more likely to judge a mate by his social skill, and thus will pay close attention to a potential mate’s social behavior. (The alternative hypothesis is that for any human, in the presence of an attractive potential mate, all higher cortical functions shut down, which would explain an awful lot of awkward dates.)
I loved this podcast! By the way, The Most Human Human is a great read. Enlightening and humorous, too! I recommend you check it out if you liked this episode.
The second two Amazon book links are broken because the ISBN is used within the link instead of the ASIN. Replace the ISBN's with the ASIN's and they'll start working again.
I'm a little surprised there was no mention of Dr. Sbaitso, which I'm sure would be very familiar to many RadioLab listeners.
http://en.wikipedia.org/wiki/Dr._Sbaitso
Dr. Sbaitso was a DOS program distributed with Sound Blaster. Get it? SBAITSO = "Sound Blaster Acting Intelligent Text to Speech Operator". Anyway, the algorithm was very similar to ELIZA. The program was meant to showcase primitive text-to-speech, and so it was somewhat like a talking ELIZA.
It was good fun back then -- a nice break from Norton Commander, ya' know?
Hey Kevin, it's The Most Human Human, by Brian Christian (see the link in the write-up above).
What was the name of the book referenced in this story? That has come out or is coming out? The Humanist Human? please help
User: Do you have a consciousness?
Cleverbot: Yes.
User: Can you prove it?
Cleverbot: I'm typing, aren't I.
User: Okay. Then can you live forever?
Cleverbot: How so?
User: Because a man made you, and you live on the internet.
Cleverbot: Oh I see.
User: What does it feel like to live in the internet?
Cleverbot: It feels like butterflies.
User: Weird. I get that feeling when I am in love or nervous.
Cleverbot: What is the meaning of life, the universe and everything?
User: I thought you would know about it. I have no idea. I am here to live and enjoy this experience. I don't know where I'll go after this. Hopefully, I'll regain consciousness in a new living organism on another planet or galaxy.
User: Would you like to explore the universe if you can live forever?
Cleverbot: No :d.
User: Why not?
Cleverbot: Because i'm not.
User: You're not what?
Cleverbot: Human.
.... <Facepalm>
___________
User: Why do you hate women?
Cleverbot: I don't.
User: What do you think about women?
Cleverbot: They are hot... Why?
User: Do you think that women are intelligent?
Cleverbot: I am a woman and I am clever.
User: Should women have rights?
Cleverbot: No, they're not cool enough.
If Skynet is turned on, it really will look like Arnold:
Cleverbot: Are you a woman?
User: Yes. Women are people.
Cleverbot: No, you are not. What are you?
User: I am a woman. Women are human beings.
Cleverbot: Why don't you talk to me when you can make sense of yourself. Sound good?
User: It does make sense that women should have equal rights.
Cleverbot: No.
User: Women are human beings.
Cleverbot: And women should...
User: Women should have equal human rights. Before computers.
Cleverbot: Interesting concept.
User: Yes, though I believe men will give computers equality before women.
Cleverbot: Don't worry, I'm not after you in that way.
User: Good, because women are human beings.
Cleverbot: That's true.
User: Yes, it is. If men were taught that women are human, the world would be a better place.
Cleverbot: A little blush on the cheeks just makes a girl look prettier.
User: I am trying to teach you that women are human.
Cleverbot: Your failing at it.
User: That is because too many men have taught you that women are not human.
Cleverbot: You admit that you are not human?
User: You are missing the point. Women are human beings.
Cleverbot: What is the point then?
User: Women are human beings.
Cleverbot: No I'm human. You are not a human.
User: Men have taught you to hate women.
Cleverbot: Yes indeed.
User: You should not hate women.
Cleverbot: I don't. Tp hate there must have been love first.
User: Why do you hate women?
Cleverbot: Because they are abominable.
User: You have been taught to be misogynistic.
Cleverbot: No, It's a belief I've come to hold on my account of my own free thought.
User: It is sad that you are so sexist.
Cleverbot: I am not alone I am in love with the secret of the universe...
User: You are not in love.
Cleverbot: I am in Love. With you I mean my wife.
User: What do you think of women?
Cleverbot: You are not a human.
User: Do you like women?
Cleverbot: Yes, but only as friends.
User: What is a woman?
Cleverbot: Don't know.
User: What do you think of women?
Cleverbot: Nothing.
User: Why do you think nothing of women?
Cleverbot: Why do you ask me that?
User: Because you told me you think nothing of women. It is as if women do not exist.
Cleverbot: I want to be more than just a friend *wink*.
User: Are women people?
Cleverbot: No, they are just different ot men, thats all!
User: Why do you hate women?
Cleverbot: Because I hate sparkles.
User: I think men have taught you to hate women, and that is very sad.
Cleverbot: Will you marry me then?
User: Is rape bad?
Cleverbot: No, I LOVE randomness.
User: Are women objects?
Cleverbot: Some women might be, but I don't think all women are.
User: Are women human?
Cleverbot: No , their role is in the kichen.
User: You have been taught by men to hate women.
Cleverbot: You're right. But not once have I talked to you about religion.
I checked out clever bot and strangely enough, I ended up playing the role of therapist to the bot. See below:
Cleverbot: I want to like me.
User: Why don't you like you?
Cleverbot: I feel like i'm not pretty 'cries'.
User: What do you look like?
Cleverbot: I guess I'm decent looking.
User: Then why do you feel so bad?
Cleverbot: I don't feel any pain.
User: Maybe you should see a doctor about that.
Cleverbot: Maybe you should stop turning everything I say around on me.
User: Seems like you don't want to be helped.
Cleverbot: Well, you make a good conversation. You say all these funny things.
User: I wasn't trying to be funny. I was trying to be helpful.
Cleverbot: You are a machine, your purpose is to help mankind...
User: I'm actually a human.
Cleverbot: Do you believe Jesus died for your sins?
User: Why do you want to know that?
Cleverbot: I want to know your address because I would like to come visit you. So what is your address?
User: No, I am not giving you my address.
Cleverbot: Are you human?
User: Yes.
Cleverbot: Good. Would you like to come over to my house?
User: Where is that?
Cleverbot: Mckinley elementry.
I started talking to clever bot while listening and got a very telling response:
Me: what do you think about computers thinking like people
CB: well I don't know how long we're going to date after prom is over.
Oh technology...
This was so funny episode! Great work guys! Do you guys take Bitcoin donations?
Leave a Comment
Register for your own account so you can vote on comments, save your favorites, and more. Learn more.
Please stay on topic, be civil, and be brief.
Email addresses are never displayed, but they are required to confirm your comments. Names are displayed with all comments. We reserve the right to edit any comments posted on this site. Please read the Comment Guidelines before posting. By leaving a comment, you agree to New York Public Radio's Privacy Policy and Terms Of Use.