Summary
Highlights
Professor David Stuckler introduces five powerful and ethical applications of ChatGPT for academic research. He highlights the dangers of misusing AI, such as plagiarism, and stresses the importance of using AI as a co-pilot and learning tool rather than a replacement for human intellect. He also mentions a free 'FastTrack AI Mentor GPT' that has been trained to turn AI usage into a learning opportunity.
The first use case involves using the FastTrack AI Mentor to help researchers, especially those struggling with broad themes like 'income inequality,' refine their research topic. The AI uses the 'convergence method' to identify broad debates, check feasibility, find personal passion links, and forecast the impact of a topic. This iterative process provides ideas and teaches models like the PICO model for further refinement, ensuring the topic is novel and impactful.
ChatGPT can significantly streamline data cleaning, particularly for quantitative analysis. The professor demonstrates how to use ChatGPT to generate code quickly for tasks like extracting years from dates or creating binary variables for missing data in Excel. This saves time by providing efficient algorithms tailored to different platforms (R, SPSS, Stata, Excel).
Professor Stuckler outlines crucial DON'Ts: 1. Do not generate full AI manuscripts and submit them to journals. 2. Do not use ChatGPT for references, as it often 'hallucinates' or fabricates them; instead, use tools like Zotero. 3. Do not let ChatGPT do the core thinking for you; use it to augment your research, not replace your own intellectual contribution.
The third use case focuses on using ChatGPT to create structured outlines, especially beneficial for narrative reviews. This helps researchers avoid getting lost in extensive literature by providing a clear roadmap for their writing. The FastTrack AI Mentor can generate outlines, define conceptual frameworks, and structure sections, allowing researchers to conduct targeted searches and organize their evidence effectively before writing.
Before using ChatGPT for editing, the professor recommends running papers through Grammarly to catch basic errors. Subsequently, the FastTrack Mentor can be used for light-touch editing, explaining the exact changes made to enhance clarity, flow, and conciseness. The key is to integrate changes selectively and transparently, declaring AI assistance to maintain academic integrity and avoid plagiarism.
ChatGPT can act as a quasi-expert peer reviewer, providing feedback from the perspective of a target journal. By prompting it to highlight strengths and weaknesses as major and minor comments, researchers can catch potential issues before submission, such as gaps in risk assessment or justification for exclusion criteria. This feature offers valuable, detailed feedback, helping researchers refine their manuscripts and turn critiques into learning opportunities.
A crucial bonus tip is using ChatGPT to develop clear research protocols. This helps researchers gain supervisor buy-in, avoid misunderstandings, and ensure ethical considerations are addressed, especially when working with large teams or sensitive data. A well-defined protocol outlines clear steps, facilitating better feedback and smoother project execution.
The video concludes by reiterating the five powerful, ethical use cases of ChatGPT for academic research and a bonus tip. Professor Stuckler emphasizes using AI as a co-pilot to speed up processing and data manipulation, not for original thinking or generating references. He encourages researchers to use the FastTrack AI Mentor as a learning and confidence-building tool, and to seek human mentorship if needed.