(Im)Paired Programming: Coding Agents Improve Productivity but Harm Understanding
Coding agents (e.g., Cursor) improve developer productivity by optimizing task completion, but shifting users from writing code to prompting and reviewing may harm their understanding, impeding oversight, learning, and communication. To probe this, we have 54 students create a website with one of two AI systems: an agent that edits user code; or a chatbot where users write code alone or adapt generic code snippets. We test understanding via comprehension questions and a task where users extend t
Record details
Published: 29 July 2026
Source: arXiv cs.HC
Category: Research
Topics: Jobs & economy · Children & education · Agents & autonomy
Retrieved: 30 July 2026
Related evidence
These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.
Open Veins of Algorithmic Auditing: Why AI Assessment Lags Behind Its Deployment in the Global South
arXiv · 23 July 2026
A Conceptual Framework for Enhancing Workforce Readiness for Smart Manufacturing in the AI Era
arXiv cs.CY · 13 August 2026
A Literature Review on New Robotics: Automation from Love to War
OpenAlex · 8 April 2015
Kenya Pushes AI and Robotics Education Through National Youth Innovation Initiative
AI for Good (ITU) · 3 August 2026
Designing Needs- and Attention-Aware AI Learning Tools for Engineering Education: Insights from Psychological Outcomes
arXiv cs.HC · 28 July 2026
Empathy and the Human-Moment Gaps of AI Chatbots: Insights from Empathy Displacement Theory
arXiv cs.CY · 29 July 2026
How to cite this record
ethics.ai (29 July 2026), “(Im)Paired Programming: Coding Agents Improve Productivity but Harm Understanding,” evidence record 14830, https://ethics.ai/record/14830 (originally published by arXiv cs.HC).
Use and limitations
This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.