"Mathematician Alan Turing proposed that machines would one day be able to think and behave like humans. This vision was challenged by neurosurgeon Geoffrey Jefferson, who argued, for example, that machines could not be classed as able to think until they had mastered language and written a sonnet.
To respond to such objections, in 1950 Turing developed a test to explore a machine’s ability to show seemingly intelligent behaviour, while hinting at his mathematical concept of imitation based on universal computing. His scientific question was whether an individual of one kind could imitate stereotypes of an individual of another kind.
Viewing Turing’s test from the perspective of benchmarks or its public misuse misses the point of his argument (B. Gonçalves The Turing Test Argument; Routledge, 2023).
Just as ideas about the meaning of the Universe were once detached from Earth, Turing sought to expand the meaning of ‘thinking’ and detach it from the anthropocentrism that contributes to the human view of both society and nature.
It is important to develop metrics for the public scrutiny of today’s generative artificial intelligence, but also to have historical perspective. We now live in one of many possible Turing futures, in which machines can pass for what they are not. Turing had good reasons to hope for some of those futures, but urged humanity to avoid others." [1]
1. What was the Turing test actually about? By Bernardo Gonçalves, Nature 624, 523 (2023)
Komentarų nėra:
Rašyti komentarą