5.3 Machines

Artificial intelligence has been a recurring theme in science fiction for a long time – since at least Samuel Butler’s 1872 novel Erewhon – whether it’s depicted as a dystopian threat to human civilisation, or a utopian ideal. But what’s the reality in the present day?

Can a computer program actually ‘think’? Can it be truly conscious? Or can it only ever simulate conscious behaviour? These are increasingly relevant questions in the 21st century. To begin to unpack them, this activity introduces you to the ‘Chinese room’ thought experiment, proposed by American philosopher John Searle in 1980.

Activity 5 The Chinese room

Timing: Allow about 10 minutes

Watch the video, and see what you make of the argument. Then consider the question beneath and make a few notes.

Download this video clip.Video player: Video 10

Show transcript|Hide transcript

Transcript: Video 10 60-Second Adventures in Thought: The Chinese Room

DAVID MITCHELL

60-Second Adventures in Thought. Number three: The Chinese Room.

[MUSIC PLAYING]

Can a machine ever be truly called intelligent? American philosopher and Rhodes scholar, John Searle certainly can. In 1980, he proposed the ‘Chinese Room’ thought experiment in order to challenge the concept of strong artificial intelligence. And not because of some 80s design fad.

He imagines himself in a room with boxes of Chinese characters he can’t understand, and a book of instructions which he can. If a Chinese speaker outside the room passes him messages under the door, Searle can follow instructions from the book to select an appropriate response. The person on the other side would think they’re chatting with a Chinese speaker, just one who doesn’t get out much. But really, it’s a confused philosopher.

Now, according to Alan Turing, the father of computer science, if a computer program can convince a human they’re communicating with another human, then it could be said to think. The Chinese Room suggests that, however well you program a computer, it doesn’t understand Chinese, it only simulates that knowledge, which isn’t really intelligence. But then, sometimes humans aren’t that intelligent either.

End transcript: Video 10 60-Second Adventures in Thought: The Chinese Room

Download

Video 10 60-Second Adventures in Thought: The Chinese Room

Interactive feature not available in single page view (see it in standard view).

Do you agree with the conclusion that a computer program could only ever ‘simulate’ intelligent thought and language comprehension? Or do you agree with Alan Turing, that a computer which can pass itself off as human (thereby passing the famous ‘Turing test’) should be said to be intelligent?

To use this interactive functionality a free OU account is required. Sign in or register.

Interactive feature not available in single page view (see it in standard view).

Discussion

This continues to be keenly debated, so there’s certainly no easy answer here. But it’s true that the philosopher in the thought experiment doesn’t understand the conversation, despite outward appearances. He’s just following step-by-step instructions. Computer programs work in a similar way – meaning a computer could engage in intelligent conversation without actually understanding it.

If you’d like to investigate the arguments further, there are some resources in the Further Reading section you might like to explore.

The focus here is the demonstration of intelligence and understanding, but this idea can be extended to the related concept of consciousness. John Searle has argued that a programmed computer model of consciousness wouldn’t actually be conscious. Others have suggested that we just couldn’t be certain about this either way. And this lack of certainty extends past machines, to a related thought experiment of sorts about ‘philosophical zombies’. This argument posits that other people around us act like normal human beings, and display the outward characteristics of being conscious. But, without getting inside their heads, we don’t really know whether they’re feeling anything, or truly experiencing the ‘qualia’ mentioned earlier. All of their responses could be just following instructions, like the computer program.