It seems to me that anything trained on internet data is kinda doomed to poison itself on the high ratio of garbage floating around here?
That same sentiment could be equally applied to humans, and not just in the internet era, but throughout all of history. There will always be misinformation and "wrong" opinions out there. "It cannot tell right from wrong" is an accusation leveled against human beings every day. We can't even all agree on what is right and wrong, truth or untruth.
A true AI is going to have to wade through all that and make its own decisions to be viewed and judged from many different perspectives, just like the rest of us.
That same sentiment could be equally applied to humans, and not just in the internet era, but throughout all of history. There will always be misinformation and "wrong" opinions out there. "It cannot tell right from wrong" is an accusation leveled against human beings every day. We can't even all agree on what is right and wrong, truth or untruth.
A true AI is going to have to wade through all that and make its own decisions to be viewed and judged from many different perspectives, just like the rest of us.