Contrary opinion: I think #7 and #8 could be feasible if the decision is left up to human judges - give them books or screenplays they've never read before, and ask them to determine which are written by AI and which are not. If one AI text is grouped together with the human-written award-winning works, then the test is passed.
Contrary opinion: I think #7 and #8 could be feasible if the decision is left up to human judges - give them books or screenplays they've never read before, and ask them to determine which are written by AI and which are not. If one AI text is grouped together with the human-written award-winning works, then the test is passed.
#9 sounds a lot more difficult, but arguably could be done with a specialized AI such as AlphaFold (the human team that developed it just won a Nobel this year for protein structure prediction).
Contrary opinion: I think #7 and #8 could be feasible if the decision is left up to human judges - give them books or screenplays they've never read before, and ask them to determine which are written by AI and which are not. If one AI text is grouped together with the human-written award-winning works, then the test is passed.
#9 sounds a lot more difficult, but arguably could be done with a specialized AI such as AlphaFold (the human team that developed it just won a Nobel this year for protein structure prediction).