The passive voice

Sometimes it’s hard to work out who is responsible.  It’s only going to get harder.

Our tutor does not normally teach English grammar, but he does handle other languages, which means he deals with such concepts as the voice of the verb.  That’s the term for whether the sentence is active (“Tom hit Dick”) or passive (“Dick was hit’); that is, whether the subject is doing something, or having something done to it.  In most writing the students are discouraged from using the passive voice, since it results generally in weaker and less meaningful prose (and in our experience, students are too much inclined to employ it).  In scientific writing it’s harder to avoid, since the emphasis is on data or images or chemicals or enzymes that are having things done to them, rather than the scientist who are doing these things.  At any rate, the main difference is that in the passive voice there is not necessarily any clue as to who or what is responsible for the action.

All other languages we’re aware of (by no means a very large number!) have passive constructions.  French and Spanish have one similar to English, a to-be verb with a participle, and in addition a reflexive: se habla español, which might be rendered as “Spanish spoken here.”  Russian has corresponding forms.  In each language it is possible, but not grammatically necessary, to add a phrase fixing responsibility: “Dick was hit by Tom.”  That can use a preposition, and we just did, or a more complex clause: “Tom, who hit Dick, ran away.”  Interestingly (to us, at least) there is another construction for fixing responsibility in Russian, in which the verb is turned into an adjective.  It can’t be formed exactly in English, but you might get some idea if in “Tom, hitter of Dick” the word “hitter” were an adjective instead of a noun.  (We do wonder why the Russians need several ways to assign responsibility.)

This linguistic diversion is, of course, irrelevant if you can’t actually assign responsibility.  Who is actually responsible for a given action or product?  Our photographer points out that this came up in a court action for copyright, when a photographer set up a situation in which a macaque took a selfie.  Who owned the copyright, the photographer or the macaque?

That’s a tiny taste of what we’re facing with the rise of Artificial Intelligence (AI), or Large Language Models (LLM), or Machine Learning (ML), to use just a few of the terms now being bandied about.  We do not claim any expertise or even special insight into these programs.  Basically, they work by training a massive set of computer nodes on input data (text, images, almost anything), and then using the resulting code to generate output.  The coders do not know how or why their output is what it is.  Much of the result is amazingly good prose, or photographically excellent images (with the occasional flat-out lie, or three-handed person).  But who is responsible for it?

We don’t have an answer.  But retreating to the passive voice won’t be enough.