COMMUNICATIONS

What we refer to as AI is given to lying, it seems. So says a historian friend of mine. For example, he asks AI about the truth of historical phenomenon that specialists have not yet explored adequately. AI responds fluently, with the confidence of one who knows what no one else has yet grasped.

But AI does not merely pretend to know all the answers. My friend grumbles at the perniciousness of it all: AI invents sources; it makes up archival material that does not exist. AI is incapable, it seems, of saying, "I don’t know; I am sorry."

How should the humanities look in the age of AI? "Figure it out," we are often told. So, why don’t we do just that? Would it not be fun to collate examples of the lies told by AI in different branches of the humanities? There should be research, I think, that analyses the tendency of AI to lie, and provides solutions to the problem.

This is not just a matter of curiosity. AI is a technique for amplifying human capabilities through computerization. Human vanity is involved here; empty rhetoric born of self-justification. AI reveals this to us in amplified form.

Is it not the role of the humanities to examine this issue, and so explore the darkness that lies within humanity itself?