The Big Picture of Researching and Analyzing with AI
The key to using AI in research is drawing the line of "what to have AI do and what humans verify." Get this wrong and you decide based on plausible lies.
The 4 Research Stages and AI's Role
- Question design: decompose a vague request into "questions to answer" (AI is good at this)
- Information gathering: collect with sources via search-specialized AI (detailed later)
- Organize/analyze: structuring, comparison, summarizing (AI is good at this)
- Verify/conclude: verify with primary sources, judge (human responsibility)
Tool Categories
- General AI: question design, organizing, analysis sparring
- Search-specialized AI: sourced latest info gathering (Perplexity, etc.)
- Material-specialized AI: answers only from materials at hand (NotebookLM, etc.)
- Agent: wholly proxy cross-research (Manus / Operator)
Absolute Principle
Don't use AI claims without sources. Always confirm figures, proper nouns, citations with primary sources. AI is a tool to speed research; humans bear responsibility for conclusions. Later chapters dig in by use and tool.
