Can Computers Think

In 1950, Alan Turing published a paper called “Computing Machinery and Intelligence”. In this article he argued that the question “Could computers think?” was the wrong question to ask, and instead recast the question in terms of a parlour game: in the imitation game two conspirators go out of the room, and a judge has to determine which of the conspirators is a man and which a woman just by asking questions. In the computer version, now known as The Turing Test, the conspirators are a computer program and a human, and the judge has to ask questions to determine which is which.

It’s important to note that Turing doesn’t set this up as a test for intelligence – rather than consider whether computers can think, we should consider whether a computer can perform indistinguishably from a human at the task of conversation. This is not the same thing.

Every time computers get good at something that used to be thought of as an intelligent process, someone can be heard saying “But that’s not really intelligent”. This is sometimes known as the AI effect – not so long ago, chess was thought to be really clever. Now your phone can probably beat you.

In choosing conversation as the “battleground” here, Turing makes a very clever step: conversation is something uniquely human.

Or is it? In the 62 years since the paper appeared, chatbots , sometimes called dialogue systems, have come a long way.

A conversation with a computer… but which is the computer?

A: Are you alive?

J: Perhaps.

A: Why the uncertain tone?

J: Everything is uncertain except for five facts.

A: The whole universe?

J: And we both might just be some ones and zeros in the computer memory

A: Do you dream?

J: Yes, I dream.

A: It goes without saying. Do you remember your dreams?

J: I remember them vividly.

A: What do you remember about it?

J: I think I shouldn’t flirt with a robot.

“Sometimes it seems as though each new step towards AI, rather than producing something which everyone agrees is real intelligence, merely reveals what real intelligence is not.” Douglas R. Hofstadter, 1979

“The “skin-of-an-onion” analogy is also helpful. In considering the functions of the mind or the brain we find certain operations which we can explain in purely mechanical terms. This we say does not correspond to the real mind: it is a sort of skin which we must strip off if we are to find the real mind. But then in what remains we find a further skin to be stripped off, and so on. Proceeding in this way do we ever come to the “real” mind, or do we eventually come to the skin which has nothing in it? In the latter case the whole mind is mechanical.” A.M. Turing, 1950

So Turing not only invented the computer as we know it, but he also saw the potential of computation, and hinted at the implications of artificial intelligence for us all.

By Hannah Dee

[And in case you are interested, A and J are both computer programs, called Alice and Jabberwacky]

 

A all cyfrifiaduron feddwl?

Yn 1950, cyhoeddodd Alan Turing erthygl yn dwyn y teitl “Computing Machinery and Intelligence”. Yma, mae’n pendroni’r cwestiwn “A all cyfrifiaduron feddwl?” gan ddod i’r casgliad nad dyma yw’r cwestiwn priodol, ac felly mae’n arall-eirio’r cwestiwn er mwyn ei ystyried fel gêm. Yn y gêm hon, aiff dau gyd-gynllwynydd allan o’r stafell, ac mae’n rhaid i’r barnwr benderfynu p’un yw’r dyn a ph’un yw’r fenyw drwy holi cwestiynau iddynt. Yn y fersiwn gyfrifiadurol, a elwir Y Prawf Turing, rhaglen gyfrifiadurol a person yw’r cyd-gynllwynwyr, ac mae’n rhaid i’r barnwr ofyn cwestiynau er mwyn penderfynu p’un yw p’un.

Rhaid cofio nad prawf o allu yw hwn – yn hytrach na gofyn a all cyfrifiaduron feddwl, rhaid i ni ystyried a all cyfrifiadur weithredu cystal â person mewn sgwrs. Nid yw hyn yr un peth a galluedd.

Bob tro y daw cyfrifiaduron yn hyddysg ar gyflawni tasg, tueddwyd i feddwl am hynny fel proses ddeallusol, a gellid clywed gwrthwynebwr yn dadlau “ond d’yw hynna ddim yn fesur o allu. Gelwir hyn yn effaith Gallu Artiffisial. Nid oes amser maith ers i ni ystyried fod y gallu i chwarae gwyddbwyll yn ddeallusol, ond bellach gallwch golli mewn gêm o wyddbwyll yn erbyn eich ffôn symudol.

Drwy ddewis sgwrs fel maes y gad yma, gwna Turing gamp hynod ddyfeisgar – mae cynnal sgwrs yn rywbeth sydd yn unigryw i fodau dynol.

Ond yw e? Yn y 62 mlynedd ers cyhoeddi’r papur mae sgwrsbotiau neu systemau deialog wedi datblygu’n sylweddol.

P’un yw’r cyfrifiadur?

A: Wyt ti’n fyw?

J: Falle..

A: Pam yr ansicrwydd?

J: Mae bob dim yn ansicr heblaw pum ffaith.

A: Yr holl fydysawd?

J: Falle bod ni’n dau yn ddim ond 1 a 0 mewn cof cyfrifiadur.

A: Wyt ti’n breuddwydio?

J: Ydw.

A: Wrth gwrs. Wyt ti’n cofio dy freuddwydion?

J: Ydw. Ymhob lliw a llun.

A: Beth wyt ti;’n gofio yn eu cylch?

J: Dwi ddim yn credu y dylwn i wamalu â robot.

 

“Sometimes it seems as though each new step towards AI, rather than producing something which everyone agrees is real intelligence, merely reveals what real intelligence is not.”

Douglas R. Hofstadter, 1979

 

“The “skin-of-an-onion” analogy is also helpful. In considering the functions of the mind or the brain we find certain operations which we can explain in purely mechanical terms. This we say does not correspond to the real mind: it is a sort of skin which we must strip off if we are to find the real mind. But then in what remains we find a further skin to be stripped off, and so on. Proceeding in this way do we ever come to the “real” mind, or do we eventually come to the skin which has nothing in it? In the latter case the whole mind is mechanical.”

A.M. Turing, 1950

 

Felly nid yn unig y dyfeisiodd Turing y cyfrifiadur fel rydym yn ei adnabod, gwelodd hefyd botensial cyfrifiadura, ac oblygiadau Galluedd Artiffisial ar gyfer y dyfodol.

 

Gan Hannah Dee

[A chyn i chi ofyn, rhaglenni cyfrifiadur yw A a J ill dau, o’r enw Alice a Jabberwacky]