@inproceedings{batra-etal-2026-catching,
title = "Catching {AI}-Generated Code at Scale: A Lightweight and Interpretable Detector for the {LLM} Era",
author = "Batra, Arnesh and
Thukral, Krish and
Sharma, Dev and
Khandelwal, Aniket and
Gumber, Arush",
editor = "Atanasova, Pepa and
Haffari, Gholamreza and
Wang, Minghan",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-industry.259/",
pages = "4357--4373",
ISBN = "979-8-89176-460-6",
abstract = "As LLMs become everyday creative partners in software development, code is no longer a purely human artefact; it emerges from an increasingly hybrid human-AI process. Understanding who contributed which parts of a codebase has become a foundational question of creative attribution, touching on authorship, trust, and the integrity of co-created work. Yet existing detectors exceed 100M$+$ parameters, are too slow for practical deployment, and offer \textit{no insight} into what distinguishes human from AI creative contributions. We present \textbf{NanoDroid}, a backbone-free detector with only \textbf{5.1M parameters} that runs $\mathbf{20\times}$ faster than 125M-parameter UniXcoder (SOTA) and is \textit{fully interpretable}: per-stream attribution scores explicitly expose which stylistic evidence (token statistics, transition patterns, surface style) distinguishes human from AI authorship in code. On the challenging 4-class DroidCollection benchmark (Human / Machine / Refined / Adversarial), NanoDroid achieves \textbf{90.77{\%}} weighted F1, within 2.6{\%} of SOTA, on the CodeT-M4 benchmark, NanoDroid achieves SOTA performance, while providing the kind of transparent decision decomposition that black-box models cannot. Beyond a detection tool, we argue NanoDroid offers a lens for studying how human and AI creativity differ stylistically, and where they blur."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="batra-etal-2026-catching">
<titleInfo>
<title>Catching AI-Generated Code at Scale: A Lightweight and Interpretable Detector for the LLM Era</title>
</titleInfo>
<name type="personal">
<namePart type="given">Arnesh</namePart>
<namePart type="family">Batra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Krish</namePart>
<namePart type="family">Thukral</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dev</namePart>
<namePart type="family">Sharma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aniket</namePart>
<namePart type="family">Khandelwal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arush</namePart>
<namePart type="family">Gumber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track</title>
</titleInfo>
<name type="personal">
<namePart type="given">Pepa</namePart>
<namePart type="family">Atanasova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gholamreza</namePart>
<namePart type="family">Haffari</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Minghan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-460-6</identifier>
</relatedItem>
<abstract>As LLMs become everyday creative partners in software development, code is no longer a purely human artefact; it emerges from an increasingly hybrid human-AI process. Understanding who contributed which parts of a codebase has become a foundational question of creative attribution, touching on authorship, trust, and the integrity of co-created work. Yet existing detectors exceed 100M+ parameters, are too slow for practical deployment, and offer no insight into what distinguishes human from AI creative contributions. We present NanoDroid, a backbone-free detector with only 5.1M parameters that runs \mathbf20\times faster than 125M-parameter UniXcoder (SOTA) and is fully interpretable: per-stream attribution scores explicitly expose which stylistic evidence (token statistics, transition patterns, surface style) distinguishes human from AI authorship in code. On the challenging 4-class DroidCollection benchmark (Human / Machine / Refined / Adversarial), NanoDroid achieves 90.77% weighted F1, within 2.6% of SOTA, on the CodeT-M4 benchmark, NanoDroid achieves SOTA performance, while providing the kind of transparent decision decomposition that black-box models cannot. Beyond a detection tool, we argue NanoDroid offers a lens for studying how human and AI creativity differ stylistically, and where they blur.</abstract>
<identifier type="citekey">batra-etal-2026-catching</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-industry.259/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>4357</start>
<end>4373</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Catching AI-Generated Code at Scale: A Lightweight and Interpretable Detector for the LLM Era
%A Batra, Arnesh
%A Thukral, Krish
%A Sharma, Dev
%A Khandelwal, Aniket
%A Gumber, Arush
%Y Atanasova, Pepa
%Y Haffari, Gholamreza
%Y Wang, Minghan
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-460-6
%F batra-etal-2026-catching
%X As LLMs become everyday creative partners in software development, code is no longer a purely human artefact; it emerges from an increasingly hybrid human-AI process. Understanding who contributed which parts of a codebase has become a foundational question of creative attribution, touching on authorship, trust, and the integrity of co-created work. Yet existing detectors exceed 100M+ parameters, are too slow for practical deployment, and offer no insight into what distinguishes human from AI creative contributions. We present NanoDroid, a backbone-free detector with only 5.1M parameters that runs \mathbf20\times faster than 125M-parameter UniXcoder (SOTA) and is fully interpretable: per-stream attribution scores explicitly expose which stylistic evidence (token statistics, transition patterns, surface style) distinguishes human from AI authorship in code. On the challenging 4-class DroidCollection benchmark (Human / Machine / Refined / Adversarial), NanoDroid achieves 90.77% weighted F1, within 2.6% of SOTA, on the CodeT-M4 benchmark, NanoDroid achieves SOTA performance, while providing the kind of transparent decision decomposition that black-box models cannot. Beyond a detection tool, we argue NanoDroid offers a lens for studying how human and AI creativity differ stylistically, and where they blur.
%U https://aclanthology.org/2026.emnlp-industry.259/
%P 4357-4373
Markdown (Informal)
[Catching AI-Generated Code at Scale: A Lightweight and Interpretable Detector for the LLM Era](https://aclanthology.org/2026.emnlp-industry.259/) (Batra et al., EMNLP 2026)
ACL