AI language models like ChatGPT should strive for political neutrality & objectivity. They certainly shouldn't pretend to be providing neutral and factual information while displaying clear political bias like ChatGPT does.
And how exactly do you see this being achieved? The fact that GPT is trained on Web data will inevitably associate a political leaning to it's output, because a large part of what's in the Internet is more or less relatable to politiccs.
Let's say that you can filter political content out of the training data (which would considerably decrease the amount of data): defining the terms is essentiel. What is even "political neutrality and objectivity"? At what point do you consider something "politically neutral"? And what do you expect GPT or any other chatbot to reply if they get asked something politically related?
That's nonsense. I think people using GPT are intelligent enough to not be influenced by what an AI can produce as an output.
5 comments
[ 4.2 ms ] story [ 19.4 ms ] threadLet's say that you can filter political content out of the training data (which would considerably decrease the amount of data): defining the terms is essentiel. What is even "political neutrality and objectivity"? At what point do you consider something "politically neutral"? And what do you expect GPT or any other chatbot to reply if they get asked something politically related?
That's nonsense. I think people using GPT are intelligent enough to not be influenced by what an AI can produce as an output.