3 ms·
I feel like the missing piece is letting AI model users define the personality they want their model to have and accepting that it can produce content we disapp
by Traubenfuchs 4y ago
I feel like the missing piece is letting AI model users define the personality they want their model to have and accepting that it can produce content we disapprove of.
Just like we don‘t hate Microsoft and Microft Word if it is used to create offensive content, we shouldn‘t hate GPT. The responsibility should lie with the user, not with the model creators or model itself.
- loa_in_ 4y agoI strongly agree. Since politics influences virtually everything, give users and let have a slider or a XY plane where one specifies expected political stance for every query. As a very nice bonus, people playfully exploring a variety of stances/worldviews/agendas(?) in connection with topics of their interest (what they queried for anyway can only possibly lead to a healthier decision making