From “Doing it All” to “Mastering the Dialogue”
Over the past year or two, AI tools—especially conversational ones like ChatGPT—have become a staple in my research workflow. Recently, I’ve noticed a fascinating shift: the time I spend on research feels less like “putting my head down and grinding” and more like “having a continuous conversation with AI.”
The Old Way vs. The AI Way
Take a recent experience of mine: I was preparing a slide deck for a conference paper:
- The Traditional Approach: I would re-read my own paper from start to finish, highlight key points, take manual notes, and then build the slides page by page. I’d painstakingly summarize the motivation, literature, methodology, and empirical results. It worked, but it usually swallowed an entire morning or afternoon.
- The New Workflow: I now feed my paper’s content into ChatGPT and ask it to generate an initial draft of the slide outline.
Is the first draft perfect? Never. Sometimes it misses the nuance, sounds too academic (even for a conference), or overlooks the specific points I want to emphasize.
The Power of the “Feedback Loop”
The real magic isn’t in the first draft; it’s in the back-and-forth dialogue. My job is no longer writing from scratch, but rather iterating with the AI:
- “Rewrite this section to be more suitable for an oral presentation.”
- “Add an illustrative example to this slide.”
- “Make the ‘Research Contribution’ section feel more practical and grounded.”
As I clarify my vision, the AI’s output aligns closer and closer to my ideal version. A task that once took half a day now takes about two hours. What’s being “saved” isn’t just typing and formatting—it’s the mental clarity gained by being forced to articulate my requirements to the AI.
The Shift in Core Competencies
This realization has changed my perspective on productivity. In the AI era, the focus of research is shifting:
It is no longer about doing every manual step yourself; it is about whether you can precisely tell the AI what you need.
The critical skill set is moving away from pure “execution” toward “workflow design, problem decomposition, and clear communication.”
Final Thoughts
The researchers who thrive in the future won’t be those who reject AI, but those who master the art of collaboration. Success will belong to those who know how to ask the right questions and possess the expertise to verify and refine the output.
AI is a powerful assistant, but the decision-making—and the name on the paper—remains ours.