{"paper":{"title":"A Large-Scale Comprehensive Measurement of AI-Generated Code in Real-World Repositories","license":"http://creativecommons.org/licenses/by/4.0/","headline":"AI-generated code in real-world repositories differs from human-written code in complexity, structure, and post-commit evolution.","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Dongfang Zhao, Haixu Tang, Hang Zhang, Tianhao Mao, Xiaofeng Wang","submitted_at":"2026-03-28T04:40:44Z","abstract_excerpt":"Large language models (LLMs) are rapidly transforming software engineering by enabling developers to generate code ranging from small snippets to entire projects. As AI-assisted code becomes increasingly integrated into real-world systems, understanding its characteristics and impact is critical. Existing study on AI-generated code is usually limited in the lab setting with synthetic benchmarks and small-scale coding tasks and covers limited metrics. AI-assisted code's manifestation in real-world codebases and its differences between human-written one remain unclear. To close this gap, we perf"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our study provides a comprehensive view of the characteristics of AI-generated code in practice and highlights how AI-assisted development differs from conventional human-driven development.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The detection pipeline that combines heuristic filtering with LLM-based classification accurately identifies AI-generated code at scale with acceptable error rates.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A large-scale study of real-world repositories finds that AI-generated code differs from human-written code in complexity, structural traits, defect indicators, and commit-level activity patterns.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"AI-generated code in real-world repositories differs from human-written code in complexity, structure, and post-commit evolution.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"e8e798e5acf199e2fd998a8ce479e8a2d3b88adc11e56494aac0254ba15336d3"},"source":{"id":"2603.27130","kind":"arxiv","version":3},"verdict":{"id":"1c88ab24-b03a-4899-b48e-b5588ee8a383","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T22:45:00.905965Z","strongest_claim":"Our study provides a comprehensive view of the characteristics of AI-generated code in practice and highlights how AI-assisted development differs from conventional human-driven development.","one_line_summary":"A large-scale study of real-world repositories finds that AI-generated code differs from human-written code in complexity, structural traits, defect indicators, and commit-level activity patterns.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The detection pipeline that combines heuristic filtering with LLM-based classification accurately identifies AI-generated code at scale with acceptable error rates.","pith_extraction_headline":"AI-generated code in real-world repositories differs from human-written code in complexity, structure, and post-commit evolution."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2603.27130/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"b93c9a9d9d948bce1d12f32a5f27b53876e0f42cb4f567b8b3f471caabbde59b"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}