Sublime

Artificial general intelligence (also sometimes referred to as “strong AI”) is still the stuff of science fiction. It is the product of a hypothetical future in which an artificial intelligence isn’t just trained to be smart about a specific task, but to learn entirely on its own, and can effectively apply its intelligence to any problem that comes

This problem requires a change in the definition of AI itself—from a field concerned with pure intelligence, independent of the objective, to a field concerned with systems that are provably beneficial for humans.

we don’t want our AI systems to go around learning every related fact individually.

This technique required not just that a giant black-box network produced accurate answers but that the individual parts could be flexibly intercombined, and all those different combinations had to work together.

I think the most worrisome aspect of AI systems in the short term is that we will give them too much autonomy without being fully aware of their limitations and vulnerabilities. We tend to anthropomorphize AI systems: we impute human qualities to them and end up overestimating the extent to which these systems can actually be fully trusted.

Artificial intelligence (AI) can already beat humans in most mind-challenging games such as chess, backgammon, quiz contests and scrabble.