AGI is a form of AI that can apply knowledge across a wide range of tasks and domains. It is an advance that moves beyond today’s artificial narrow intelligence (ANI), which can generally only function somewhat accurately, in most cases within a specific set of parameters.
AGI is expected to be able to mimic humans’ cognitive capabilities to learn, understand abstract concepts, reason and carry out complicated tasks, at times even moving beyond humans’ capacity to identify and solve complex problems due to its speed and knowledge capacity.
Can nimble and reliable AGI be innovated, designed, tested and put into wide use globally by 2040? All told, 231 of the 328 experts and analysts who responded to at least one question in this Fall 2023 canvassing answered a set of quantitative questions about the likely advancement of effective AI with human-like general intelligence by 2040.
They were not asked to write an open-ended answer, so no written qualitative data specific to AGI was gathered in this canvassing.
The respondents’ opinions on all three AGI questions were a nearly even split. The results:
Question 1: Will there be scientific consensus by 2040 that AGI has been achieved?
– 49% said it is very unlikely or somewhat unlikely
– 45% said it is very likely or somewhat likely
– 6% said they don’t know.
Question 2: How do the advances you see in AGI make you feel?
– 39% are equally excited and concerned
– 37% said they are more concerned than excited
– 17% are more excited than concerned
– 7% are neither excited nor concerned.
Question 3: How likely do you think it is that AGI could pose an existential risk to humanity at some point in the future, probably beyond 2040?
– 48% think it is very or somewhat likely such a risk could be posed
– 47% believe it is very or somewhat unlikely
– 5% say they don’t know.
Continue reading: Research methodologies and topline findings