Two analogies for AI and ethics
16 Oct 2026I’m trying to get some opinion on the typical conversation that goes like this:
– I don’t want to use AI, nobody should use AI.
– You have no choice. Soon it will not make sense not to use it.
In this post I explore two analogies related to this situation, without any conclusion.
Analogies can be misleading, and some might think reasoning should be abstract, not example-based. To me they are useful. In my mind it is similar to the following situation. If you tell me that I can get a coffee but there is a probability of 10^{-10} that it is poisoned, I will hesitate because I cannot refer to a similar situation; but if you tell me that this probability is much lower than the probability of being hit by a meteorite during my lifetime, then I will drink it.
Analogy 1: Computation by hand
Consider some computation that is done by hand, and can then be replaced by computers. For example, some approximately solving partial differential equations for dam safety.
Imagine that we are at the moment where some people use computers for this task, and others don’t. People who do not use computers might have several reasons for not doing so: they cannot afford a computer, they are against using computers made by unethical companies, they value the process of doing these computations and the community around it.
The likely scenario here is that the community converges on using computers for this task. At some point, the reviewers will consider that computation by hand has no value, because most people in the community can do the same in a second on their computers. And for the remaining people doing computation by hand, then either they get a job in a museum that preserves this know-how, or change the way they present their research (“low-tech alternatives”?), or move on to doing something completely different.
There is some period when computation by hand is still valued, either because the reviewing peers do not have computers yet, or because there is some inertia or a feeling of empathy.
All the work that have been done without computers will have some weight on a CV for a few years, but this weight decreases with time.
Analogy 2: Torturing monkeys
Consider a situation where a research group obtains groundbreaking results of the highest technical quality, but tortured thousands of monkeys in the process (think about something even worse, if needed).
How does the community adapt to this situation? Let’s assume that almost everybody refuses to torture monkeys.
Are people going to pretend this experiment never happened and they have never seen the results? I guess not. Even if people wanted to, they will hesitate to spend years of their lives trying to contradict something that everyone secretly knows to be established. Maybe people would try to reproduce the results in a more acceptable way, in an effort to restart from a clean basis. And then the community as a whole would consider this to be the work to be praised, I guess?