Abstract
<jats:p>As artificial intelligence tools become increasingly integrated into software development workflows, they offer significant gains in productivity but simultaneously introduce new challenges for code transparency and auditability. While AI-assisted coding accelerates task completion, reduces boilerplate effort, and lowers barriers to entry for novice developers, it also obscures the provenance of code contributions, complicates code review processes, and creates difficulties in tracing the origin of bugs or security vulnerabilities. This research systematically assesses the trade-offs between maintaining high levels of transparency in version-controlled repositories and preserving the workflow efficiency gains that motivate AI adoption in the first place. Through a mixed-methods approach combining controlled experiments, developer surveys, and repository analysis, the study evaluates how different transparency-enforcing mechanisms—such as mandatory AI labeling, prompt logging, and deterministic attribution—impact developer productivity, code quality, and team collaboration. The findings contribute actionable guidelines for development teams seeking to balance the benefits of AI acceleration with the need for accountable, maintainable, and trustworthy codebases.</jats:p>