Using gpt-3 for personality and/or worldview modelling?

Hi, I’m just wondering whether it might be viable to use gpt-3 to analyse personality and worldview based on response to certain concepts. For example, the user might rank certain words based on their response to them, give one word responses to simple questions, or even input word associations. I’m sure gpt-3 might even be able to give an analysis of the answers to a Rorschach test if given enough sample data, although it might be best to be used in conjunction with a different model. Anyway I would like to know if this is a feasible idea, and if it is, what might be the best way of doing it?

This is almost exactly what @Kopaka is working on. @DR.AMES

1 Like