What is the Turing Test?

0:00 Welcome to Philosophy is not boring by kidopoly.com. I'm Mira. And I'm Finn! If you haven't already, tap follow wherever you're listening so you never miss an episode. Okay Finn, I want you to picture this. Imagine you were in your room, texting your best friend on a tablet. You were talking about your favourite video game, you were laughing and making jokes.

0:24 Okay, sounds totally normal so far.

0:27 But what if, on the other end of those messages, there is no human being at all? What if you were actually just texting a big, glowing box of wires?

0:37 Like a robot pretending to be my friend? That is incredibly creepy. I would definitely figure it out. Would you really? How? I'd ask it something a robot would never know. Like what it feels like to scrape your knee, or if it is ever scared of the dark. A robot doesn't have knees or fear.

0:57 But what if the robot gives you the absolute perfect answer? What if it texts back, yeah, the dark is terrifying. I always make sure my mum leaves a nightlight on. If a machine can talk exactly like a human, is it just tricking you? Or is it actually thinking?

1:14 Whoa, wait. Can a computer actually think? That is exactly the massive question we are exploring today. But Mira, what does thinking even mean? My calculator can do math way faster than I can, but I don't think it's thinking. It's just doing math.

1:31 That is a brilliant point, Finn. People have argued for centuries about what the word think actually means. Is it having feelings? Is it being smart? Is it having a soul?

1:46 Yeah, because a machine doesn't have a brain or a heart. It's just a bunch of metal and computer chips.

1:52 Exactly. And back in the year 1950, a very famous mathematician named Alan Turing was chewing on this exact problem. He was working at a university in Manchester, in England.

2:06 Wait, 1950? Computers back then were gigantic. I saw a picture once. They took up entire rooms and had thousands of messy wires.

2:17 They totally did. They used things called vacuum tubes. They hummed really loudly, and they got incredibly hot. It was hard to imagine something that looked like a giant heater actually having a thought.

2:31 So did Alan Turing think those giant heaters could think.

2:35 Well, Turing was a genius. He looked at those clunky metal machines and imagined a future where they could be super smart. But he realized that arguing about the definition of the word think was just getting everybody stuck. So, he said, let's stop arguing about what thinking means. Let's try playing a game instead.

2:57 Ooh, I love games. What kind of game? Like hide and seek? Kind of.

3:03 He called it the imitation game. But today, almost everyone calls it the Turing test. And it's basically a test to see if a computer can hide in plain sight. The Turing test. Okay, how do you play it? Picture three completely separate rooms. You are in room A. Hidden in room B is a real human being, and hidden in room C is a computer.

3:30 You can't see them and you can't hear their voices. You can only type text messages to them and they type messages back.

3:38 So I'm just passing notes under the door to two strangers.

3:43 Yes, you chat with both of them, and your job as the judge is to figure out which room has the human and which room has the machine.

3:52 Oh, I would totally win this. I would ask a really tricky question. I'd ask. What is the most dangerous animal you've ever had to fight?

4:03 Okay, great question. The human in room B types back, a really angry goose at the park when I was five.

4:12 Geese are absolutely terrifying. Good answer. What does the computer say?

4:18 The computer searches its massive database of millions of human stories in a split second and types back, once a huge hairy spider dropped on my head and I screamed so loud my neighbour came over.

4:32 Wait, a computer doesn't have a head? Or a neighbour. It is completely lying.

4:38 Yes, it is pretending. That is the entire point of the game. Turing said, if you play this game for a long time and you truly cannot tell which one is the machine, then you have to admit the machine is intelligent.

4:53 But it is just copying us. It doesn't actually know what a spider feels like.

4:59 Ah, but Alan Turing's argument was, who cares? If it talks like it thinks and acts like it thinks, isn't that proof enough?

5:09 I don't know about that. Tricking someone just means you are good at lying. It isn't the same as actually being smart inside your own head.

5:17 Finn, 30 years later, in 1980, an American philosopher named John Searle completely agreed with you.

5:25 See? I knew I was onto something. What did John Searle say? He came up with his own famous thought experiment to prove Alan Turing wrong. It is called the Chinese room. The Chinese room? Does someone get locked in a room again? Yes. Imagine we lock you in a room, and someone slides questions under the door to you.

5:47 But the questions are all written in Chinese characters. But I don't speak a single word of Chinese. I wouldn't understand anything on the paper.

5:56 I know, but inside the room with you is a giant, massive English rule book. The book says, if someone gives you this specific curvy squiggle, copy down this other spiky squoggle.

6:09 Oh, so I don't need to know what the words mean. I just look at the shape of the symbol, find it in the book, and draw the shape the book tells me to draw.

6:18 Exactly! You match the shapes perfectly and slide your answers back under the door. Now, outside the door, there are native Chinese speakers reading your answers.

6:29 Wait, do my answers actually make sense to them?

6:32 They make perfect sense. Your rulebook is so good that you are telling brilliant jokes and answering deep questions in Chinese. The people outside think there is a fluent, intelligent Chinese speaker in the room.

6:45 But there isn't. It's just me, and I have absolutely no idea what I'm saying. Boom!

6:52 Sal said that is exactly what a computer does. A computer doesn't understand anything. It is just following a giant code rulebook, matching shapes and spitting out answers without feeling a single thing.

7:05 Whoa! So a computer is just a guy in a room with a rulebook, doing it super fast.

7:12 That's what John Cell believed. He called the rulebook syntax, which means the rules of the code. But he said computers have no semantics, which means actual meaning or understanding.

7:24 That makes sense. But Mira, why does this matter to us right now?

7:29 Because right now, in your real life, chatbots can easily pass the Turing test. AI programs write essays, they tell jokes, and they talk to you just like humans.

7:40 Yeah, sometimes when I ask a smart speaker a question, it sounds so real I forget it's a robot.

7:46 And that's why this matters. Next time you use a chatbot to help with homework or ask a smart speaker a question, you can ask yourself, is anybody actually home in there? Should we trust machines to make big decisions for us if they don't actually understand what they are saying?

8:04 It's like they are really good at faking it, but the lights are on and nobody is home. Here's a mind-bending fun fact.

8:12 In the 1960s, a programmer built the very first chatbot named Eliza. It was designed to talk like a doctor. Even though people knew it was a simple machine, they still told it all their deepest secrets. We just really want to connect. By the way, Alan Turing, who started all this, was a true hero. He helped crack secret codes to end World War II faster.

8:40 Sadly, he was treated very terribly later in his life by his government just for who he loved. But today, he is celebrated as a genius who basically gave birth to computer science. Alright, Finn, it is time for the quiz. Are you ready?

8:57 Okay, my brain is full of robots and rule books, but I'm ready.

9:01 First question. What was the name of the mathematician who invented the imitation game in 1950? Next one. Question three. In 1980, John Searle came up with a famous thought experiment to prove computers don't understand things. What was the name of the room? Here's number four. In that thought experiment, what language were the questions written in?

9:46 Last question. What was the name of the very first chatbot in the 1960s that pretended to be a doctor? Great job! Let's see how you did. The first answer is Alan Cheering. For question two, it's called the Cheering Test. Number three, the answer is the Chinese room. Question four. The questions were written in Chinese.

10:17 And the final answer is Eliza.

10:20 Phew, I'm glad I didn't have to use a giant rule book to answer that quiz. Me too. What are you going to do now? I'm going to go interrogate the smart speaker in the kitchen and see if I can trick it into admitting it's a robot.

10:35 Good luck with that. If you had fun wondering about robots, minds, and chatbots with us today, please leave us a five-star review. Just scroll down and tap the stars.

10:46 Yeah, and tap follow on this show wherever you're listening, and check out our sister shows. History and science is not boring, and maths is not boring. Go give those a listen and a follow as well.

10:58 Want your name on the show or just want to say hi? Send us an email at hello at kidopoly.com.

11:06 See you next time on Philosophy is not boring. Bye everyone.