When I was a graduate student in artificial intelligence, the humanities were not held in high regard. They were vague and woolly, they employed impenetrable jargons, and they engaged in “meta-level bickering that never decides anything”. Although my teachers and fellow students were almost unanimous in their contempt for the social sciences, several of them (not all, but many) were moved to apoplexy by philosophy. Periodically they would convene impromptu Two-Minute Hate sessions to compare notes on the arrogance and futility of philosophy and its claims on the territory of AI research. “They’ve had two thousand years and look what they’ve accomplished. Now it’s our turn.” “Anything that you can’t explain in five minutes probably isn’t worth knowing.” They distinguished between “just talking” and “doing”, where “doing” meant proving mathematical theorems and writing computer programs. A new graduate student in our laboratory, hearing of my interest in philosophy, once sat me down and asked in all seriousness, “Is it true that you don’t actually do anything, that you just say how things are?” It was not, in fact, true, but I felt with great force the threat of ostracism implicit in the notion that I was “not doing any real work”.

Source: Philip E. Agre, “The Soul Gained and Lost: Artificial Intelligence as a Philosophical Project,” Stanford Humanities Review 4, no. 2 (1995): 1–19.

Do computer scientists really not critique or interpret the structures, assumptions, and ideas informing their code and calculations?

Philip Agre was trained as an AI researcher before becoming one of the field’s most powerful internal critics. So, I don’t know how many grains of salt I should bring to his account. As a humanist, I find the passage funny, unsettling, maybe even strangely useful (with a good bit of eye-rolling). I also don’t know if I am supposed to care about what “AI people,” as Agre likes to call them, think about the humanities. But I also wonder what is being claimed here. Do computer scientists really not critique or interpret the structures, assumptions, and ideas informing their code and calculations? If they don’t, I can certainly teach them how to. Lol. My guess is that Agre is describing a particular culture within AI at a particular historical moment. Perhaps things have changed since 1995? Still, I find the statement interest for all the inter-disciplinary drama it brings 🙂