notebookLM did miracles for my research paper
Last semester, I crammed my research paper, and it got rejected. They told me that my topic was broad enough, enough to generate two paper and told me to actually write two papers in the span of 3 WEEKS. I am not joking.
I used to use ChatGPT but realized that if you’re not careful, it’s got a tendency to generate non-existent author names, like who is Bonnie Jackson bro?
My friend suggested to use NotebookLM and bro was a life-saver. But you still need to put in some work.
Here’s what I did;
1.) Dumped my actual sources into NotebookLM.
Instead of asking it random questions, I uploaded the journal articles, PDFs, reports, and other sources I was actually allowed to use. This kept my research grounded in the materials I had.
2.) Used it to interrogate my sources.
I asked things like “What are the main arguments across these studies?”, “Where do these authors disagree?”, and “Which findings are relevant to my research question?” It saved me from rereading 30 PDFs just to find one tiny detail.
3.) Made it help me find gaps, not invent information.
I uploaded my draft paper and asked it to identify relevant backup data I can use to cite my claims, recurring themes, contradictions, limitations, and areas that the existing studies weren't addressing. That’s where my actual research direction started to become clearer.
4.) Made an outline before writing.
Once I had enough information, I used the findings to organize my RRL and research sections before touching the actual paragraphs. This stopped me from writing 1,500 words and realizing none of them belonged there.
5.) STILL CHECKED EVERYTHING.
NotebookLM made research faster, not foolproof. I still opened the original papers, checked the citations, verified the claims, and wrote the actual paper myself.
Honestly, the biggest change was going from “AI, please write my research paper” to “Here are my sources. Help me understand what the hell I’m looking at.”