13 Enhancing clarity
Latifah Alsalem, Postgraduate Research Student, Computing, United Kingdom
During my PhD research, I regularly analysed published papers to underpin my theoretical framework. Whenever I encountered complex code snippets, I turned to generative AI to clarify syntax, annotate functions, and compare alternative model architectures. In the literature review phase, I used them to summarise large volumes of text, identify key themes, and draft initial paragraphs. As I progressed to empirical work, the same tools helped me draft and revise sections of my reports, check grammar, and refine my writing style. I requested worked examples and step-by-step solutions for mathematical derivations and statistical calculations, which I then adapted to my datasets. Additionally, I used these tools to troubleshoot LaTeX errors in Overleaf, correcting syntax for figures, tables, and bibliography entries without interrupting my writing flow. Overall, I integrated these capabilities at every stage of my research journey, from early exploratory reading and coding to finalising journal-style reports and ensuring consistency, precision, and coherence throughout the project.
I relied chiefly on three platforms: ChatGPT, Perplexity and SciSpace. When writing code, I iterated prompts such as “Add inline comments to this script” until I obtained clear, contextual annotations. To deepen my understanding of my codebase, I asked for step-by-step explanations of complex functions and for a high-level overview of the entire workflow, which helped me see how individual modules interconnect. I also pasted snippets of my Overleaf source to request corrections to LaTeX syntax, from fixing misplaced braces to resolving package conflicts, and then applied the revisions directly in my project. To check my drafts, I sought grammar and style feedback, then compared suggestions across Perplexity and ChatGPT to choose the most fitting formulation. During the literature review, I uploaded PDFs to SciSpace to extract summaries of key arguments and to generate citation suggestions. This iterative workflow, prompt, review, and refine became integral to my study routine and greatly improved both my comprehension and the quality of my outputs.
… generating multiple versions of paragraph openings or section headings helped me overcome writer’s block
Firstly, they saved a significant amount of time, reducing hours of manual reading or Google searches to mere minutes. For instance, brainstorming research questions and organising literary themes became effortless once I provided a topic and asked for thematic outlines. Secondly, these tools enhanced clarity: when I struggled with dense theoretical passages, concise paraphrases illuminated underlying concepts, allowing me to grasp complex ideas more rapidly. Thirdly, they streamlined document preparation: resolving Overleaf syntax errors and generating LaTeX snippets for tables and figures in seconds meant I could focus on content rather than formatting. Finally, they supported creativity and structure: generating multiple versions of paragraph openings or section headings helped me overcome writer’s block and establish a coherent narrative flow. While I always cross-checked my technical accuracy, the overall effect was a more efficient, organised, and confident approach to both coding and academic writing.
GenAI Tool(s) Used:
