THÔNG TIN TÀI LIỆU
A my K ind can think Here I will focus on this last element of Turing’s work,6 but all four of these elements are part and parcel of what has sometimes been called the computer revolution – a revolution that parallels previous scientific revolutions associated with figures such as Copernicus, Darwin, and Freud As Luciano Floridi has argued, the computer revolution has forced us to reconceptualize the world in which we live: Computer science and the issuing technological applications have not only provided unprecedented epistemic and engineering powers over natural and artificial realities; by doing so they have also cast new light on who we are, how we are related to the world and hence how we understand ourselves (Floridi 2012, 3540) When approaching the question of whether machines can think, it is easy to get bogged down in definitional matters such as what we mean by “machine” and what we mean by “thinking” – so much so that the very question itself may seem a meaningless one To bypass this difficulty, Turing proposed that we instead approach the question by judging the machine’s performance in an imitation game Consider a set-up involving a neutral interviewer and two systems – one human and one machine The interviewer asks questions of each system, not knowing which is which, via some kind of remote relay (Turing described this in terms of a teleprinter connecting the rooms, but today it is easy to imagine the set-up in terms of texting or some other kind of instant messaging.) With no limit placed on the type of questions allowed, the interviewer is then given some set amount of time – five minutes, say – and is charged with determining which system is the human and which system is the machine Having described this imitation game, Turing suggests we discard the meaningless question “Can machines think?” and instead ask: “Are there imaginable machines that can well at the imitation game” (Turing 1950, 442)? Turing himself gave an affirmative answer to this question, and he famously predicted that by the end of the 20th century, machines would be developed that would be able to reliably succeed at the imitation game While this prediction has not yet come true, machines have exhibited considerable mastery across a diverse range of seemingly intelligent domains Not only have machines been able to beat humans at games such as chess, Go, and Jeopardy!, but we now also have unmanned drones in our skies, driverless cars about to populate our highways, and programs like Siri and Cortana that organize our lives.7 To many, it seems only a matter of time – perhaps only a very short matter of time – before the Turing test is passed (See Chapter 12 for discussion of the likely future of artificial intelligence.) However the future unfolds, it’s hard to overstate the importance played in the late 20th century by the framework for thinking about machine intelligence that Turing’s imitation game provided In particular, this framework was almost solely
Ngày đăng: 29/10/2022, 21:16
Xem thêm: