New Research: AI Involvement and Code Drift: Comparison of Drift Density and Severity by Provenance

Mon, 12 Oct 2026 00:35:18 +0000

We just posted the write-up of our most recent research. Here's the abstract:

AI Involvement and Code Drift: A Provenance-labeled Comparison of Drift Density and Severity Across Nine Production-readiness Dimensions

Abstract
BACKGROUND: Prior work on drift in AI-generated code has largely measured freshly generated components in isolation, leaving open the question of what AI involvement does to real, shipping codebases. Because fresh generation (greenfield) and mature production (brownfield) code differ in age, review history, and accumulated debt, a like-for-like comparison requires real code on both sides of the ledger. This study samples repositories labeled by degree of AI involvement in advance, and scans them on identical instrumentation, so that drift is compared across provenance groups on real code, rather than between freshly generated and mature corpora.

METHODS: Twenty-one open-source repositories, selected by GitHub search rules with six added by hand, were labeled by authorship provenance before scanning into three groups: human-led (n=6), AI-assisted (human + AI; n=10), or AI-led (n=5). Inclusion in an AI group means that AI was used. Full or partial authorship could not be guaranteed. Using ReWeaver, a multi-dimensional deterministic code drift scanning engine, we scanned repositories at the application root (a full walk of up to 2,000 files per scan root). Analysis included whole-repository Production Drift Ratio (PDR; a measure of drift frequency weighted by remediation time) scores, and line-level drift density, as well as breakdowns on nine dimensions of production readiness: user experience, design consistency, accessibility, security & privacy, reliability, architecture, maintainability, testability, and AI code governance.

RESULTS: Both PDR and drift density rose monotonically with the degree of AI involvement (M PDR: human-led = 0.773; AI-assisted = 0.830; AI-led = 0.835). Drift density followed a similar pattern (human-led = 4.42%; AI-assisted = 6.39%; AI-led = 10.15%). Added drift in AI-led code concentrated in design-facing dimensions: User Experience (+0.34 PDR), Accessibility (+0.17), and Design Consistency (+0.10). Maintainability, the largest single source of drift in the human-led baseline, was roughly equivalent in PDR between tiers.

CONCLUSION: On real code compared on identical instrumentation, drift was found to grow with AI involvement, specifically where design intent lives.

Takeaways:

Drift was present in every repository at every provenance tier. No authorship category, including mature, human-led production code, produced drift-free software. Both drift metrics increased
monotonically with AI involvement.

Whole-repository PDR (max score=1.0) rose from 0.773 to 0.835, and drift density from 4.42% to 10.15% moving from human-led to AI-led repositories.

AI-led repositories carried roughly 2.3× the drift density of the human-led baseline.

Added drift concentrated in design-facing dimensions: User Experience (+0.34 PDR, 6.1× human density), Accessibility (+0.17, 2.3×), and Design Consistency (+0.10, 2.6×).

It should be noted that provenance labels are not guaranteed. While the independent authorship detector agreed with all of the human-led designations, it disagreed with 14 of the 15 AI group labels and was inconclusive on the fifteenth.

Results should be interpreted as a signal for future research.

Find more about the study and download the paper here: http://www.reweaver.ai/provenance-study

Coder Legion - Recent posts tagged drift

New Research: AI Involvement and Code Drift: Comparison of Drift Density and Severity by Provenance