You can, but ChatGPT as a specific product is trained to be biased, it's been shown many times. OpenAI isn't unbiased so nor are their products.

Such things should be based on actual open source LLMs like the LLaMA models we can tune directly and audit openly.