Amerika

The great thing about computer science is that we are working with reality modeling. Computers are mathematically precise, and as a result, if given the right input data, they can reveal to us some of how reality operates, like any other scientific experiment:

Interestingly, this extends to lying AIs who act just like humans:

When incorrect answers were flagged, getting better at giving correct answers was one way to optimize things. The problem was getting better at hiding incompetence worked just as well. Human supervisors simply didn’t flag wrong answers that appeared good and coherent enough to them.

In other words, if a human didn’t know whether an answer was correct, they wouldn’t be able to penalize wrong but convincing-sounding answers.

The AIs lie because we told them that doing so was rewarding. One key question is when and how often do we get lied to.

When your average human is asked a question he does not understand and is in a situation where “I don’t know” is a bad answer, he will make some stuff up. For example, if a cartel leader asks one of his mules where the drugs went, the mule is going to invent a story about being robbed to cover his own neck.

As it turns out, AI does the same thing because it is the logically correct thing to do. Unfortunately it has no idea that its lies are transparent. This mirrors the Dunning-Kruger Effect in humans, where people cannot understand anything above their intellectual level and make absurd choices instead.

Tags: ai, dunning-kruger effect, incompetence, lying