ChatGPT was designed to assume information that is not given. This is what the documentation says:
“Ideally the model would ask clarifying questions when the user provided an ambiguous query. Instead, our models usually guess what the user intended.”
If you do not want it to guess or make up information then I recommend adding something like the following to your prompts:
“Do not make up any data, statistics or facts. Instead, please ask clarifying questions for anything that isn’t clear”
You may get more questions back than you expect because human language can be pretty ambiguous to a machine. Adding this to your prompts is usually a good starting point though and it will often reduce the amount of fact-checking that you have to do afterwards.
Comments