Inspire Digital: Could your smartphone think?

Does your smartphone think? Is it conscious? Hopefully not. But could there ever be a smartphone, or any computer, that thinks with a rich, conscious, mental life of comparable complexity to your own? In this article Professor Nicholas Jones considers some of the answers to these questions…

This question might seem utterly intractable. It’s natural to regard our own consciousness as something private to each of us. You have your thoughts and experiences. I have mine. Neither of us can access the other’s. I can’t think your thoughts, or have your experiences, just as you can’t mine. What goes for other humans goes for smartphones too. So even if smartphones could think, we could never know, because we could never think their thoughts or have their experiences.

Yet that can’t be the whole story. Somehow or other, we do all know that other human beings have conscious inner lives. You know that if you step on my toe, I will feel pain. And I know the same about you. We know this because we interact with other people and they respond: ouch, that hurts, get off my toe! We learn to interpret these responses and thereby learn how other people think. This knowledge is imperfect – it’s incomplete and sometimes wrong – but it’s real. It’s also central to our relationships with other people. We treat other conscious people very differently from unthinking stones.

That’s how we know that others think. But our question wasn’t about knowing whether smartphones think. Our question was whether they do or could think, regardless of whether we knew. Those are different questions. Yet they were ingeniously connected by Alan Turing, a founder of theoretical computing. His proposal is now called the Turing Test.

Here’s a version of the Test. You’re having two text message conversations. You can ask questions in each conversation. Your conversational partner has to respond. One of them is an ordinary human. The other is a computer. The computer passes the Test if you can’t tell which is which. Turing’s idea: if the computer can reliably pass the test, then it thinks. To put it slightly differently and more generally: to think is just to speak and act how we ordinarily take thinkers to speak and act.

Is that right? A thought experiment proposed by Ned Block suggests not. Imagine a mad scientist who knows how you will respond to every possible input. They program a machine that looks just like you to respond exactly as you would, in every possible circumstance. Unlike you, the machine doesn’t use complex neural processing to decide how to act. Instead it just consults the mad scientist’s list of possible inputs and corresponding responses. Since you would pass Turing’s Test, this machine would too. Yet the machine isn’t thinking, Block claimed, just consulting a database.

Should we accept that Block’s machine isn’t thinking? It’s processing and responding to information. Why does it matter just how the machine does so? And if Turing’s Test is no good, our knowledge of other people is thrown into doubt. The idea behind the Test is just that the way we usually decide what others are thinking is a good way of doing so. If something could pass the Test without really thinking, our ordinary ways of deciding whether others think are inconclusive too. How could you then know that you’re not the only conscious thinker, in an otherwise thought-free universe?

Professor Nicholas Jones is Tutorial Fellow in Philosophy at St. John’s College, Oxford

Challenge

Do you really know that other human beings have conscious thought?

If so, how do you know this? And could you also know in this way whether a computer had conscious thought?
If not, should that affect how you behave towards other people? For example, is it ok to stamp on other people’s toes, given that you don’t know pain will result?

Is Block’s machine really possible? How does it differ from human beings? And how, if at all, does that matter to Turing’s Test?

CLCIK HERE TO SUBMIT YOUR WORK!