Not at all, there's not even a "being" there to have those opinions. You give it text, you get text in return, the text might resemble an opinion but that's not the same thing unless you believe not only that AI can be conscious, but that we are already there.
“What are you doing?”, asked Minsky.
“I am training a randomly wired neural net to play Tic-Tac-Toe” Sussman replied.
“Why is the net wired randomly?”, asked Minsky.
“I do not want it to have any preconceptions of how to play”, Sussman said.
Minsky then shut his eyes.
“Why do you close your eyes?”, Sussman asked his teacher.
“So that the room will be empty.”
At that moment, Sussman was enlightened.
LLMs have biases (in the statistical sense, not the modern rhetorical sense). They don’t have opinions or goals or aspirations.
Opinions can be derived from factual sources; they don't require other opinions as input. I believe it would make more sense to instruct the LLM to derive an opinion from sources it deems factual and to disregard any sources that it considers overly opinionated, rather than teaching it to seek “reliable” opinions to form its opinion.