Accountability in Code Review: The Role of Intrinsic Drivers and the Impact of LLMs
Bibliographic record
Abstract
Accountability is an innate part of social systems. It maintains stability and ensures positive pressure on individuals’ decision-making. As actors in a social system, software developers are accountable to their team and organization for their decisions. However, the drivers of accountability and how it changes behavior in software development are less understood. In this study, we look at how the social aspects of code review affect software engineers’ sense of accountability for code quality. Since Software Engineering (SE) is increasingly involving Large Language Models (LLM) assistance, we also evaluate the impact on accountability when introducing LLM-assisted code reviews. We carried out a two-phased sequential qualitative study ( \(\textbf{interviews}\rightarrow\textbf{focus groups}\) ). In Phase I (16 interviews), we sought to investigate the intrinsic drivers of software engineers influencing their sense of accountability for code quality, relying on self-reported claims. In Phase II, we tested these traits in a more natural setting by simulating traditional peer-led reviews with focus groups and then LLM-assisted review sessions. We found that there are four key intrinsic drivers of accountability for code quality: personal standards , professional integrity , pride in code quality , and maintaining one’s reputation . In a traditional peer-led review, we observed a transition from individual to collective accountability when code reviews are initiated. We also found that the introduction of LLM-assisted reviews disrupts this accountability process, challenging the reciprocity of accountability taking place in peer-led evaluations, i.e., one cannot be accountable to an LLM. Our findings imply that the introduction of AI into SE must preserve social integrity and collective accountability mechanisms.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.055 | 0.273 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.002 |
| Science and technology studies | 0.005 | 0.010 |
| Scholarly communication | 0.010 | 0.009 |
| Open science | 0.002 | 0.008 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".