Combating Harms of Generative AI in CS1 with Code Review Interviews and a Flipped Classroom

Peter Fowles; Erik Falor; Sulove Bhattarai; John Edwards; Seth Poulsen

Home/Archive/arXiv

arXiv · 2605.21374

Combating Harms of Generative AI in CS1 with Code Review Interviews and a Flipped Classroom

Peter Fowles·Erik Falor·Sulove Bhattarai·John Edwards·Seth Poulsen

View original source ↗Read source document ↗Export metadata ↓

Abstract

Background and Context: Large Language Models (LLMs) are more accessible and accurate than ever before, raising significant concerns for computing educators. One major concern is students using LLMs to bypass the effort needed to understand concepts and metacognitive strategies essential for success in computer science. Objectives: We contribute a unique approach to assessing and building up student understanding through weekly oral code review assessments. These formative assessments incentivize students to understand their submitted code, regardless of whether or not the code was generated by AI tools. We also use a flipped classroom to provide time for students to learn concepts outside of class and provide ample time for students to schedule code review interviews. Methods: For this paper, we collected data from three semesters. We analyze student exam scores, keystroke logs, and surveys to understand how the new course policies affected student learning, behavior, and attitudes. Findings: Pairwise comparison of exam results reveals a statistically insignificant increase in average scores for Fall 2025 compared to previous semesters. Keystroke logs show a significant increase in characters pasted per total characters input into coding assignments in Fall 2025, pointing towards higher AI usage. Survey results show positive student sentiment towards code reviews at the end of Fall 2025, with nearly all negative feedback being addressable through better scheduling and more rigorous TA training. Implications: Oral code reviews with a flipped classroom appear to be effective at mitigating harms of LLM use while providing space for students to freely experiment with these tools. Our work suggests that students in Fall 2025 still show adequate understanding of material covered in written exams, despite dramatic increases in LLM usage for coding assignments.

Explore related subjects

cs.HC

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Peter Fowles, Erik Falor, Sulove Bhattarai, John Edwards, Seth Poulsen. 2026-05-20. Combating Harms of Generative AI in CS1 with Code Review Interviews and a Flipped Classroom. https://arxiv.org/abs/2605.21374

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Explanation Navigator: Rectifying Out-of-Scope Human Interpretations of Leaky AI Explanations through Conversational Guidance

As explanations of artificial intelligence systems proliferate, their recipients must grasp not only what they convey but also recognise what they cannot. We conducted an interview study with nine participants to examine how explainees reason when their information needs exceed the scope of available explanations. Participants often unwittingly confabulated explanatory insights when relevant information was missing from the explanations, not recognising the inherent limitations thereof. We characterise such explanations as leaky explanations -- simplifications that strive to hide complexity yet whose correct interpretation hinges on understanding of the concealed details. To address out-of-scope interpretations we propose Explanation Navigator: a conversational interaction framework that detects mismatches between users' information needs and explanations' content, elucidating pertinent yet implicit details and providing complementary explanations for unmet information needs. An online study with 316 participants showed that our approach allowed explainees to recognise and rectify confabulated explanatory insights, guiding them towards developing correct understanding.

cs.HC↗

arXiv2026

Biased AI improves human performance but reduces perceived helpfulness

Artificial intelligence (AI) increasingly shapes how people think, engage, and evaluate information. To minimize risk, most current systems are designed to present as ideologically neutral with standardized output. Yet growing evidence suggests that these principles suppress cognitive engagement, impair human decision-making, and erode societal diversity. Here we test the opposite approach by deliberately injecting bias into AI assistants. In three randomized experiments with 5,000 participants, biased AI improved human performance relative to default and neutral AI in tasks ranging from misinformation evaluation and financial investment to graduate education. These gains carried a subjective cost. Participants systematically undervalued AI they believed to be biased and inflated the helpfulness of AI they believed to be neutral, regardless of the systems' actual behavior. Interacting with two AIs whose biases flanked the participant's own perspective preserved the performance gains while limiting the subjective cost and one-sided influence. Our findings reveal the strategic value of intentional bias in AI design. Rather than performing a single fair, reliable, and authoritative voice, AI that speaks from specific viewpoints triggers cognitive agency and elevates human-AI performance in judgment, decision-making, and problem-solving.

cs.HC↗

arXiv2026

Learning Password Best Practices Through In-Task Instruction

Users often make security- and privacy-relevant decisions without a clear understanding of the rules that govern safe behavior. We introduce pedagogical friction, a design approach that inserts brief, instructional interactions at the moment of action. We evaluate this approach in the context of password creation, a familiar task with clear quality criteria. We conducted a randomized study with 128 participants across four interface conditions that varied the depth and interactivity of guidance. We assessed three outcomes: (1) rule compliance in a subsequent password task without guidance, (2) accuracy on survey questions tied to password rules, and (3) behavior-knowledge alignment, which captures whether participants who correctly followed a rule also recognized it on the survey. Across the guided conditions, participants corrected most rule violations in the follow-up task and showed high behavior-knowledge alignment. Survey results suggested clearer advantages for some rule types, especially symbol related questions. These results position pedagogical friction as a lightweight intervention for security- and privacy-critical interfaces.

cs.HC↗