Beyond Machine Authorship: Visible Account Provenance and the Integration of Human, Bot, and AI-Agent Pull Requests

Under review. A large-scale study of public GitHub pull requests showing that the visible account a change is published under — a human login, an AI agent under a developer login, or an AI agent under an autonomous app account — tracks how often the change is merged, with autonomous app-account agent PRs merging least often.

Authors: Iman YeckehZaare · Venue/status: ACM Collective Intelligence

This submitted manuscript analyzes three public GitHub datasets totaling tens of thousands of merged and agent-authored pull requests and finds that AI-agent pull requests merge less often than human PRs, least of all under an autonomous app account, even after controlling for task type and repository popularity. Task-context, same-issue, line-reuse, CI, downstream-revert, and authorship-masked LLM-adjudication checks all leave the association in place, and screened maintainer comments show no explicit anti-AI policy — pointing to quiet, distributed closure rather than stated rules.

Public artifact boundary: this route exposes status, authorship, visual summary, citation metadata, and the contribution boundary; manuscript files are posted only when review and prepublication rules allow it.

The rendered manuscript page adds status, visual summary, review boundary, citation metadata, and contribution notes. Key links: home, systems, papers, manuscripts, Google Scholar, GitHub, LinkedIn, ORCID, MIT profile, and CV PDF.