Modern ChatGPT Workspaces and Multimodal Tools
Lesson 2: Search, Deep Research, and Source Verification
Lesson Objectives
By the end of this lesson, students should be able to:
- Choose ordinary chat, web search, or deep research based on the task
- Define a research question, source scope, and deliverable
- Evaluate citations for relevance, authority, recency, and claim support
- Separate sourced findings from inference and recommendation
Lesson Content
Use the lightest research mode that meets the stakes. Ordinary chat is useful for stable concepts and drafting. Search is appropriate for current facts, prices, schedules, rules, product details, or when links matter. Deep research is appropriate when the task requires multiple sources, comparison, synthesis, and a documented report.
A research request should define: the decision or question, date boundary, geography or jurisdiction, preferred primary sources, excluded sources, comparison criteria, and required output. Ask the system to label verified facts, disputed claims, missing evidence, and its own inferences.
Citation inspection is mandatory. A citation is not proof merely because it exists. Open the source. Confirm that it supports the exact sentence, is current enough, represents the correct jurisdiction, and is authoritative for the claim. For legal, medical, financial, safety, or business-critical decisions, use licensed or accountable human review.
Research stop rule: stop when additional sources no longer change the conclusion, major viewpoints have been represented, and the remaining uncertainty is explicit. More links do not automatically mean better research.
Enroll to continue this lesson.
The preview above shows the lesson objectives and opening lesson content. Enroll to view the full lesson, complete the practice work, and take the lesson quiz.
