As technology develops, do you think it will ever make sense to say that a computer "knows" things?

That depends, of course, on what you mean by "know".

On one well-known, though flawed, definition, to know that P is to have a justified true belief that P. Suppose that you are willing to say that a computer believes-true a proposition P if P is true and the computer has (a representation of) P stored in its database. And suppose that a computer would be justified in such a belief if it deduced it from axioms (also stored in its database). Because some computers can deduce propositions according to rules of logic, such a computer could, on this definition of "know", know that P.

Sometimes, people feel that to know something is also to be aware that one knows it, or that one knows that one knows it. So, could a computer know something in this sense? That, of course, depends on what might be meant by "aware". But it is certainly possible for a computer to believe that it believes something (for example, in the sense that it not only has P stored in its database, but also has something like "I have P stored in my database" stored in its database). So it's surely plausible that such a computer could be "aware" that it knows that P.

My colleagues and I investigated some of these possibilities with a computer system that could represent and reason about beliefs, including its own beliefs. One of our papers on this topic is:

Rapaport, William J.; Shapiro, Stuart C.; & Wiebe, Janyce M. (1997), "Quasi-Indexicals and Knowledge Reports", Cognitive Science 21: 63-107.

online here.

Read another response by William Rapaport
Read another response about Knowledge