On bad questions (series: notes to myself)

Some questions are bad. Not because they are stupid. Stupid questions are so because of their content, or context, or implications. Stupid questions can also be bad, of course, but bad questions need not be stupid. Questions are bad because they make you look bad when you answer them. For they are questions badly formulated. They are "loaded". So if you wish to avoid a silly answer, you must be ready to question the question, and this is not only bad manners, it also gives the impression that one is trying to dodge the question.

Bad questions seem to be of at least two kinds.

Yes or no questions. "So, in the end, and to be clear about what you really think: is true, universal Artificial Intelligence possible or not?". If you say no, you are wrong, because there is a sense in which AI is possible, i.e. it is not logically impossible (a contradiction). But if you say yes, then you are wrong, because true, universal AI, the kind you see in sci-fi movies, is factually impossible, in the ordinary sense in which it is impossible for a fair coin to land always on the same side. It won't, even if that possibility cannot be logically excluded. So a good answer starts with "it depends on what you mean by "possible"...", and this is really bad manners. Because it is not an answer, but a way of questioning the question. At a conference, I was once asked: "Can a computer think?", I started answering by saying "it depends on what you mean by "think"...", but I never finished the sentence. Everybody thought it was such a funny joke, they all enjoyed it very much, and so we moved on to the next question... Later people told me I had been very witty. I never had the courage to tell the audience I seriously meant that it really depends on what you mean by "think".

Either-or questions. Same story but with no contradictory statements, just inconsistent ones. "So, in the end, and to be clear about what you really think: is AI a good or bad technology?" Of course, the answer is neither, or rather both, good for something and bad for something else. So any absolute answer is silly. But if you start asking "in what context and in what sense...", then that is dismissed as bad manners.

Bad questions can be avoided, of course, but you have seen the catch. Those who avoid them look bad. Like they do not want to commit themselves to a straight and honest answer, a clear cut yes or no, good or bad. Hair-splitting people, who, like politicians, will not tell you what they really think. 

Bad questions are asked by incompetent people or journalists in search of a one-line quote. If you do not give them the required answer (they usually do not care which one, as long as it is only one) they will not be satisfied (plenty of interviews where the nuanced answer did not make it to the printed text, because the journalist needed a yes or a no, a good or a bad, not a real understanding of the issue). I recall a pre-program interview during which a BBC journalist asked me a yes/no answer. I told her that things were more complicated. She admitted that she knew it perfectly well, but the audience needed to hear two clear and opposite positions, one for the yes and one for the no, and then enjoy the debate. Nuances did not grab the attention of the audience, she said. She needed a sort of duel. I insisted that I could not see the value in two positions being so polarised and both wrong; that this was a disservice (especially by the BBC); that surely the public was able to appreciate slightly less childish yes/no and more mature, qualified answers. She agreed, but also told me that that was the format, and as I was not ready to play that game, I never took part in the program. Her last question was whether I could recommend a colleague who was ready to go public with a yes or a no answer. I declined. I know several of them, but the last thing they need is more public exposure. 

In academia, bad questions are used by so-called "practical people", those who start their questions with some falsy-modest claim about them being only concrete, realistic, practical, pragmatic people. "Let's be clear, I am a concrete person ... yes or no?". "You know me, I am very pragmatic: ... good or bad?".

I have met many academics who think that this level of bluntness is a virtue. Like people thinking that an axe is so much better than a scalpel. At the first distinction, as soon as you start introducing a clarification, they huff and puff, impatient and annoyed by what they perceive as a lack of simple, honest, genuine commitment to a clear and straightforward answer. "Can't you make up your mind, for goodness' sake? We are not doing philosophy here we are ... [feel free to fill the gap]". 

Next time a bad question is asked, instead of feeling bad about questioning it, better make the person asking it feel bad about it, by showing that it is a poorly formulated question.