Download word docx Download MP3 Listen to MP3 online
Read below for script, questions and answers
The Turing Test
Hi. Good afternoon. Today I want to talk to you about something called the Turing Test. The Turing Test was invented by a man called Alan Turing. He lived from 1912 till 1954. Maybe you’ve heard of him. Maybe not. If you haven’t heard of him, you have definitely used something that he invented, or part invented. The computer. Have you ever used a computer? If you have, then you’ve worked with something that he invented.
Before the Second World War, Alan Turing was a mathematician. He specialized in code. He was called a cryptographer. During the Second World War, the British army very much wanted to read the code of the German army. However, the German army used a machine called an Enigma machine, which made the code extremely difficult to read. Alan Turing, and a few other geniuses, were hired by the British government to try and crack these codes. They couldn’t do it through pure brain power, so Alan Turing came up with a machine that cracked the codes for him. This machine, kind of a calculator, in the end, became the modern computer.
Now, when Alan Turing was working on these things, he came up with something that he called the Turing Test. The Turing Test is a test of computer intelligence. It’s a test to see if a computer can fool a human into believing that the computer is a real person.
The Turing Test works like this. You have a questioner and you have a person and you have a computer. The questioner cannot see the person and cannot see the computer, obviously. The questioner asks questions to both the person and the computer by typing on a console. There is no voice. You cannot hear them. The questioner asks the questions and the person and the computer reply to the questions. The questioner can ask any type of question at all. The challenge … the point of the test is to see if the questioner can guess which is the person and which is the computer. If the questioner guesses correctly, the computer fails the Turing Test. If the questioner guesses incorrectly, if the computer is able to fool the person into thinking they are a real human, then the computer wins the Turing Test. Alan Turing thought that this would be a good way of testing computer intelligence. To see if computer intelligence could approximate, or impersonate, human intelligence.
Now, the Turing Test has been beaten. It was beaten years ago. In fact, if you have an iPhone, Siri on your iPhone, could probably pass this test. If you ask Siri questions, Siri will answer you in … er… normal human language. And it sounds almost human. In fact, you can have a conversation with Siri. There are websites online that you can go to and you can input questions and an AI system, Artificial Intelligence system, will reply to your questions and it’s so realistic that you would be forgiven for thinking you were talking to a real person.
So, computers have reached the point where they can beat the Turing Test pretty easily. However, the point of Alan Turing’s test was to see if computers could persuade another human that they are a human. But, that doesn’t necessarily mean the computers have intelligence. Siri, on the iPhone, as I just talked about, can impersonate intelligence. Siri can pretend to be intelligent by having a conversation, but is Siri actually thinking? If I ask Siri a question, is Siri actually thinking? Or is Siri merely going through a few steps of programming to find the answer for me.
Now, the interesting thing about artificial intelligence and the way we are going at the moment is will computers be able to think? And if they can, how do we tell that they are thinking? This gets very very confusing and very very difficult to think about. If a computer can answer my questions, is the computer thinking or is the computer approximating thinking? And, if the computer is actually thinking, how do we know? How do we know the computer hasn’t just been programmed to give those responses? How does the computer know that? How do I know that I am thinking? How do I know that I have free rein over my thoughts? How do I know that I can say whatever I want? How do I know I haven’t been preprogrammed? Banana. That was a surprising thing to say. Could I say that because I have no programming? Or could I say that because I have programming? How do we know we are intelligent? How do we know we think? How do computers know they think? How do we know computers think? How can we ever quantify… how can we ever understand intelligence like this? It’s a very very difficult question. If Alan Turing was alive today, maybe he would come up with another test. I don’t know, but he would certainly find it interesting. Anyway, thank you for listening, I think. Or do I think? I don’t know.
The Turing Test – Questions
- How old was Alan Turing when he died?
A: 12
B: 19
C: 42
D: 54
- What do you call a person who studies and makes codes?
A: A mathematician
B: A cryptographer
C: A genius
D: An enigma
- Why was Alan Turing hired during World War 2?
A: To make codes for the British army
B: To crack British codes
C: To make codes for the British army
D: To crack German codes
- Why did Alan Turing invent a machine to crack codes?
A: They couldn’t do it simply by using their brains.
B: They couldn’t find where the German codes were.
C: They didn’t have enough geniuses working together.
D: They liked to think that a computer was a real person.
- Why would Siri be able to pass the Turing Test?
A: It is a system on the iPhone.
B: It can answer your questions.
C: It appears almost human.
D: It can understand normal human language.
- What is the point of the Turing Test?
A: To answer questions that the questioner asks
B: To persuade a human that a computer is human
C: To show that computers can understand intelligence
D: To show that computers have artificial intelligence
- Steven says, “How do I know that I have free rein over my thoughts?” What do you think “free rein” means?
A: Freedom
B: Weather
C: Power
D: Control
- What is Steven’s main idea?
A: That we will never find a computer that can really beat the Turing Test
B: That we can’t know if computers are really intelligent
C: That computers in the future will control us
D: That AI will bring many advantages to our lives
- In the Turing Test, why do you think the questioner, the computer and the human type their questions and answers?
- Draw a diagram to explain the Turing Test.
- Why do you think Steven suddenly says, “banana”? What conclusion does he draw from this?
- Can we ever know that we think?
- AI will be good for humankind. Do you agree or disagree?
The Turing Test – Answers
- C 2. B 3. D 4. A 5. C 6. B 7. A 8. B
- In the Turing Test, why do you think the questioner, the computer and the human type their questions and answers?
They have to type their questions because if it was a verbal test, the computer would be found out instantly. Computers can mimic human language almost perfectly, but they cannot yet mimic pronunciation, intonation, inflection and feeling. Think about Siri. It sounds very good, but you would never mistake it for a human because it sounds … off, for want of a better word.
- Draw a diagram to explain the Turing Test.
- Why do you think Steven suddenly says, “banana”? What conclusion does he draw from this?
He says, “banana,” because he is trying to demonstrate that he can say something unexpected. He is presumably thinking that a computer would only be able to follow pre-established lines of conversation and would not be able to say something unconnected and surprising. However, he then goes on to draw the conclusion that this doesn’t necessarily prove anything. An AI system could be programmed to occasionally say surprising and unconnected things. Being able to say something that seems unpredictable doesn’t mean it is unpredictable. It doesn’t prove free will, it just proves that you can say it.
- Can we ever know that we think?
The philosopher Descartes said, “I think, therefore I am.” He assumed that the only thing we could truly know about ourselves and our surroundings is the fact that we think about them. Therefore, we must exist because we can think we exist. However, he was philosophizing a long time before the invention of AI machines. If he wrote again, now, he might have to change what he believes. I think it is now possible to think without being aware or intelligent. A computer can “think” in that it processes programming to come to a solution. A computer could be programmed to day dream or hope for the future, but it wouldn’t mean that the computer was really thinking in the way that we think. But, then, the problem is how do we know that we actually do think? How do we know that we are not merely AI machines who have been programmed to simulate thinking? The answer is we don’t, and we never can, unless we have a moment of awakening like Neo in the Matrix movie. So, the answer is no. And, in the end, it is unhelpful to think about. We can never know the answer, so it doesn’t benefit us to think about it. We think, therefore we are. Probably.
- AI will be good for humankind. Do you agree or disagree?
I think AI could be good for humankind if it is dealt with well. There are three problems that I think need to be considered. These are whether we control the AI, whether the AI is used for good or evil, and what we will do in an AI society.
The first problem is whether we will be able to control the AI or whether it will control us. At the present time, we have self-learning AI systems that improve themselves. These systems are extremely task specific. For example, Google’s AlphaGo AI system was designed to win at the game of Go. It has beaten most of the major grandmasters in the world, and is probably unbeatable. It played itself millions of times and learned from its experience. It taught itself to get better. However, this is all that the system can do. It cannot drive a car, deliver a parcel or hack into a military database. The software and algorithms may be useful in designing future AI, but right now the system is completely within our control. Without humans, this software would not be able to exist. The problem may arise when machine intelligence becomes too advanced for us. At that point, we will need AI systems to design and build AI systems. When that happens, we will have effectively taken ourselves out of the loop and the machines will be in charge of the machines, and, by extension, us. It is still early days and this will only happen if we allow it to happen. Checks can be put in place to limit it.
The second consideration is the use that AI systems are put to. We could make systems that spend their time analyzing and curing diseases. Or we could make autonomous killing drones for the military. Both would use AI. One would benefit humankind, the other wouldn’t. Simplistically speaking, good or evil. How this problem ends up will depend entirely on where the research money comes from. If history is anything to go by, the military have a larger research budget than most private corporations.
The third thing we have to start thinking about is what will happen to us in a society with high functioning AI. Presumably, most of the jobs we currently do will have been taken over by AI. In fact, Donald Trump’s call to “make America great again” is largely hindered by the fact that these jobs have been lost to automation and not China, as he claimed. This process will only increase. So, what will humans do? That is something we need to spend a lot of time considering. With careful planning and well thought out welfare programs, it could be an extremely positive thing. The machines will do our jobs and we can enjoy creation, art, human interactions and so on. We need the governments to then spread the generated wealth around so that we can continue to live comfortably. If they don’t. If the 1% continue to get richer and we lose our jobs at the same time, then the world will be a troubled place.
So, I see AI as having a positive impact on humankind. It may be an optimistic view, but I see a world where advanced AI systems have cured all known diseases, solved climate change and given us the free time to enjoy life as humans should. It could happen.