Re:An Ethical AI Never Says “I”
by 8note on Sunday March 26, 2023 @06:17PM
(#35320232)
"I" belongs in LLMs because "I" shows up in the training set. Plenty of text is written from the first person perspective, and for auto-complete tasks, the first person is useful
Eg. You task the ai with rewriting a letter for you. The resulting text will include plenty of "I"s, and it should.
Eg. You task the ai with rewriting a letter for you. The resulting text will include plenty of "I"s, and it should.