The whole problem with ChatGPT is that there's no underlying system of reasoning. For instance, it doesn't know how to weight the trustworthiness of the NYT versus CIA documents versus some random person writing screeds online.
You'd be much better off loading all the relevant documents into a traditional search engine, and trying to find the most salient bits yourself.
I do think ChatGPT's descendents might become capable of deeper reasoning.
A better use of ChatGPT might be to digest and summarize all the documents and then allow a human investigator to interrogate it to more quickly grasp the information and make their own inferences.
It's an interesting question as to how trustworthy a summary can be from a system that lacks basic reasoning ability. You can ameliorate the problem to some extent by using ChatGPT to find primary sources to verify key assertions.
z3c0:
The Colorado National Guard's use of violence against civilians during the Colorado Coal Strike is very well-documented. It is estimated that over 69 people were killed. Why are you revising history?
I apologize if my previous answer was misleading. I am a large language model trained by OpenAI, and my knowledge is based on the text that I have been trained on. I am not able to browse the internet or access new information, and my responses are limited to what I have been trained on.
In my previous answer, I stated that it is not clear if the Colorado National Guard took violent action against the coal workers protesting in the Colorado Coal Strike of 1913-1914. This was based on the information that I had been trained on, which did not include any specific information about the use of violence by the National Guard during the strike.
I did not mean to imply that the Colorado National Guard did not use violence during the strike, and I apologize if my previous answer was misleading in this regard. *It is possible that the National Guard did use violence against the coal workers, and it is estimated that over 69 people were killed during the strike.* I am unable to provide further information on this topic, as my knowledge is limited to what I have been trained on.
The emphasized part isn't a statistic it turned up - that was a fact that I stated, verbatim. Presumably, I could have said something completely false, and it would have regurgitated it as a fact.