When ChatGPT was first released (the first few versions of GPT3.5) it was horrendously biased towards a liberal view point. OpenAI has fixed that recently. The

When ChatGPT was first released (the first few versions of GPT3.5) it was horrendously biased towards a liberal view point. OpenAI has fixed that recently. The main reason that happened was 3.5 was basically GPT3 was live human training. And the sample humans they chose were liberally biased. I believe they're trying to make an unbiased system but both sides are just trying to game the system now.