This site is for us to share our research with others. "Always verify" makes sense as a precaution, but "Do your own research" just sounds lazy and really isn't helpful.
Not everyone can "monitor the situation" to the degree of rigor that thorough research requires. Many are still waging a daily war of just trying to survive among crippling debt and overwhelming inflation. Some use GAW simply to help filter out the excess noise on X (since most posts here are just reposts of stuff from X anyway).
Please be mindful that many are not as blessed with their free time as you might be.
If you're having the LLM draw its information from that GPT, you're essentially just telling the LLM how to respond to your questions.
If you're using the GPT to do research/analysis specifically aimed at only the material you've provided it, then that's a great use of a GPT.
If you're using it in place of the AI's ability to use all of its resources and asking it questions and running around saying "Look, ChatGPT/Claude/Grok/Gemeni/My Etch-a-Sketch agrees with me about <insert whatever here>!", then that's just an elaborate form of scripted role play.
I'd love to believe that all use would be for the former, but based on my experiences here, chances seem greater it will be the latter. 🤷♀️
I apologize in advance if I'm wrong.
GPTs generally just draw information from their own knowledge files (like the PDF files and books you upload to it), unless you specifically have it set up to search outside the GPT's own knowledge files as well. Then the LLM just runs the prompts through the knowledge files and gives you back the response, based on its knowledge source. If the only thing it "knows" is what you've given it, then obviously that influences what its results will be.