# How Scientists Use Large Language Models to Program

arXiv: https://arxiv.org/abs/2502.17348 · HTML: https://mlsanity.com/paper/2502.17348/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Feb 2025.

## Built on by (most cited first, 8 of 8)

- [Unified Architecture Metamodel of Information Systems Developed by Generative AI](https://mlsanity.com/paper/2604.00171/) (2026)
- [More Than "Means to an End": Supporting Reasoning with Transparently Designed AI Data Science Processes](https://mlsanity.com/paper/2603.24877/) (2026)
- [User Misconceptions of LLM-Based Conversational Programming Assistants](https://mlsanity.com/paper/2510.25662/) (2025)
- [What happens when reviewers receive AI feedback in their reviews?](https://mlsanity.com/paper/2602.13817/) (2026)
- [Beyond the Desk: Barriers and Future Opportunities for AI to Assist Scientists in Embodied Physical Tasks](https://mlsanity.com/paper/2603.19504/) (2026)
- [IndustryCode: A Benchmark for Industry Code Generation](https://mlsanity.com/paper/2604.02729/) (2026)
- [LLMoxie: Exploring Agentic AI for Scientific Software Development](https://mlsanity.com/paper/2607.02703/) (2026)
- [Aleena: Alignment Agent for Research Software Engineering Collaborations](https://mlsanity.com/paper/2607.08043/) (2026)

## Abstract (arXiv)

Scientists across disciplines write code for critical activities like data collection and generation, statistical modeling, and visualization. As large language models that can generate code have become widely available, scientists may increasingly use these models during research software development. We investigate the characteristics of scientists who are early-adopters of code generating models and conduct interviews with scientists at a public, research-focused university. Through interviews and reviews of user interaction logs, we see that scientists often use code generating models as an information retrieval tool for navigating unfamiliar programming languages and libraries. We present findings about their verification strategies and discuss potential vulnerabilities that may emerge from code generation practices unknowingly influencing the parameters of scientific analyses.