The Bias in AI
Jonathan Cloud January 30th, 2024
The more I play with ChatGPT, the more it seems to me to have a bias — a bias toward the conventional, toward an incorrectly “balanced” view, as well as a superficial tone and an attraction for clichés and bad writing.
As an example of what I’m talking about, consider the Chatbot’s response to Jonathan Rowson’s recent provocation on war and peace:
“Do we have a theory and practice of peace that is worthy of the risks and challenges of the 21st century? If not, how do we expect to survive? And if so, what would that look like?”[1]