Rebooting AI: Building Artificial Intelligence We Can Trust

gullibility gap, which starts with the fact that we humans did not evolve to distinguish between humans and machines—which leaves us easily fooled.

Seventh, the combination of preexisting societal biases and the echo effect can lead to the amplification of social bias.

Third, modern machine learning depends heavily on the precise details of large training sets, and such systems often break if they are applied to new problems that extend beyond the particular data sets on which they were trained.

human beings who themselves base their actions on abstractions like ideas, beliefs, and desires.

A big part of why deep learning struggles in language, for example, is that in language, new sentences with fresh meanings are in infinite supply, each one subtly different from the last.

neural networks can’t give human-style explanations for their answers, correct or otherwise.*7

COGNITION MAKES EXTENSIVE USE OF INTERNAL REPRESENTATIONS.

The largest effort in the field, by far, is a project known as CYC, directed by Doug Lenat,

Similar techniques can be used to represent part/whole relationships.

common sense: a rich understanding of the world, and how it works, and what can and cannot plausibly happen in various circumstances.