Augmentation with Dilution: A Large-Scale Empirical Study of Human Contributor Ecosystems After AI Coding Agent Adoption

2026-02-24 · Source: cs.SE updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Data Science & Analytics · Depth: Expert, extended

Summary

A large-scale empirical study, "Augmentation with Dilution," analyzed 11,097 GitHub repositories from January 2023 to May 2026 using a staggered difference-in-differences design to assess the impact of AI coding agent adoption on human contributor ecosystems. The research found no significant change in the absolute number of human contributors ($ATT=0.014$, $p=0.224$), but a significant 1.9 percentage point reduction in human contributor density ($ATT=-0.019$, $p=0.002$), indicating a declining relative share of human participation. Newcomer relative participation share decreased significantly by 3.7 percentage points ($ATT=-0.037$, $p<0.001$), an effect emerging immediately and remaining stable. Additionally, review depth increased by 5.3% ($ATT=+0.0168$, $p<0.001$), suggesting a shift of burden from code production to review. These effects vary by project size, programming language (Python, TypeScript), and maturity, collectively demonstrating "augmentation with dilution" in open-source ecosystems.

Key takeaway

For Directors of AI/ML overseeing open-source contributions, recognize that AI coding agents, while augmenting code production, dilute human participation and increase review overhead. Your teams should proactively implement strategies to attract and retain newcomers, especially in smaller or younger Python/TypeScript projects, and allocate increased resources for code review, particularly in larger, more mature projects, to mitigate potential long-term sustainability risks.

Key insights

AI coding agents dilute human participation and increase review burden in open-source projects without displacing contributors.

Principles

• AI agents reshape open-source participation structures.
• Newcomer share and review burden are key project health metrics.
• Project characteristics moderate AI agent impact.

Method

Staggered difference-in-differences design with Sun and Abraham estimator on GitHub data, using propensity score matching for control group construction and repository/month fixed effects.

In practice

• Monitor human contributor density in projects.
• Prioritize newcomer engagement strategies.
• Increase code review resource allocation.

Topics

• AI Coding Agents
• Open-Source Software
• Human Contributor Ecosystems
• Causal Inference
• Code Review
• Newcomer Participation

Best for: CTO, VP of Engineering/Data, AI Scientist, Research Scientist, Director of AI/ML

Related on AIssential

• Agentic Coding and the Economics of Open Source — AI-driven "vibe coding" shifts software economics from open source to personalized, on-demand development, impacting developer incentives and attention.
• Mining Architectural Quality Under Agentic AI Adoption: A Causal Study of Java Repositories — AI coding tool adoption does not inherently improve software architecture quality; density metrics can be misleading.
• It’s the Lessons We Learned Along the Way. Or, Is It? — AI agents can efficiently complete complex data tasks, but may diminish human understanding and critical thinking development.
• Claude Opus 4.6 and GPT-5.3-Codex 30 Mins Apart — AI agent teams are rapidly advancing, demonstrating autonomous complex task execution and self-improvement capabilities.
• Code Is Not Cheap: How to Multiply Your AI’s Output With Software Fundamentals — AI-assisted coding, while prevalent, demands human engineering expertise to mitigate quality issues from "jagged" model outputs.
• ‘Tokenmaxxing’ is making developers less productive than they think — Increased AI code generation correlates with disproportionately higher code churn and revision rates, undermining productivity claims.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.SE updates on arXiv.org.