tool for extracting moral beliefs

riceissa commented 5 years ago

I'm imagining some sort of interactive tool where there is a series of moral situations, like trolley problems, pascal's wager, lifespan dilemma, etc., and the user is asked what outcome they prefer, and they keep doing this for many dilemmas, and then the output you get is you see what your "abstract moral beliefs" are, or if they are incoherent, you get to see which inputs the user gave conflict (and where they conflict, maybe the user gets to resolve them or maybe some other set of dilemmas is presented to help you resolve them). So in this way you get to "build up" your values using concrete cases. And if many people use this tool, you get to learn a lot about moral beliefs.

You might call this "guided/automated reflective equilibrium". This seems to basically be what philosophers already do manually by reading lots of books, papers, and blog posts that go over these ethical dilemmas.

Does something like this already exist?

Actually https://aiwatch.issarice.com/?organization=EthicsNet seems a little similar (but they seem to have a different purpose?)

vipulnaik commented 5 years ago

There is the moral foundations questionnaire done by Jonathan Haidt et al. See the "Moral foundations" section at https://vipulnaik.com/moral-profile/

riceissa commented 3 years ago

I think the spencer greenberg cluster has now produced one or more things like this, but I haven't looked at it yet.

riceissa / project-ideas

tool for extracting moral beliefs #62