Should ChatGPT be Biased? Challenges and Risks of Bias in Large Language Models