Ask HN: ChatGPT and Product Liability
This, as far as I know, is a domain that has not been tested in the courts.
The scenario would start with a product liability lawsuit, say, for a piece of industrial machinery, self driving car, aircraft, etc.
During discovery the lawyers learn that code snippets where generated using AI coding tools. Yes, the code was analyzed by a human and tested. However, court cases are not necessarily decided on the basis of logic or scientific analysis. All you have to do is convince twelve people of something, anything, that will get you a guilty verdict.
That's where we face the reality of a general population that is deficient in the requisite knowledge and understanding as well as all the fear mongering permeating the AI conversation today.
I posit it would not be too difficult to convince a jury that the AI-generated code is evidence enough of negligence and skewer the accused with a guilty decision.
I would think that if they can identify a single section of AI-generated code they might be able to convince the jury that there is no way to know just how much of the codebase was generated using these tools (which could be a fair assumption).
Thoughts?
Any attorneys with legal opinion on this?