---
format: "aidr-story-markdown/v1"
id: "873b10b62bbef9da343091273353a52441a4bc3cd1148d30e977b0055b145ef1"
canonical_url: "https://aidr.today/873b10b6?lang=en"
title: "Human vs Machine Translation Detection: A Cross-Model, Cross-Domain, and Low-Resource Analysis"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-17T05:00:00.000Z"
category: "Research"
topics: ["llm"]
source_urls: ["https://arxiv.org/abs/2611.18586"]
summary: "The growing reliance on web-crawled corpora for training large language models introduces a critical challenge: machine-translated (MT) content is often treated as human translation (HT), potentially degrading data quality and propagating synthetic artifacts, especially in low-resource settings. In this work, we study MT-HT detectability across high-resource (English, Spanish) and low-resource (Swahili, Afrikaans) languages using both decoder-only LLMs in zero-shot and few-shot settings, and fine-tuned multilingual encoders trained at the sentence level and on token-chunked inputs. We find that decoder-only models provide limited discrimination overall and exhibit a consistent class-asymmetric failure mode: they recognize human-written text reliably (high HT recall) but rarely flag fluent machine output (low MT recall). Encoder-based classifiers achieve high accuracy when evaluated on data generated by the same MT systems and translators used during training in a legislative domain, but their MT recall drops sharply on unseen translators and on unseen domains such as low-resource conversational text and medical content."
---

# Human vs Machine Translation Detection: A Cross\-Model, Cross\-Domain, and Low\-Resource Analysis

> [Open the canonical story](<https://aidr.today/873b10b6?lang=en>)

**Published:** 2026-11-17T05:00:00.000Z
**Category:** Research
**Topics:** llm

## Summary

The growing reliance on web\-crawled corpora for training large language models introduces a critical challenge: machine\-translated \(MT\) content is often treated as human translation \(HT\), potentially degrading data quality and propagating synthetic artifacts, especially in low\-resource settings\. In this work, we study MT\-HT detectability across high\-resource \(English, Spanish\) and low\-resource \(Swahili, Afrikaans\) languages using both decoder\-only LLMs in zero\-shot and few\-shot settings, and fine\-tuned multilingual encoders trained at the sentence level and on token\-chunked inputs\. We find that decoder\-only models provide limited discrimination overall and exhibit a consistent class\-asymmetric failure mode: they recognize human\-written text reliably \(high HT recall\) but rarely flag fluent machine output \(low MT recall\)\. Encoder\-based classifiers achieve high accuracy when evaluated on data generated by the same MT systems and translators used during training in a legislative domain, but their MT recall drops sharply on unseen translators and on unseen domains such as low\-resource conversational text and medical content\.

## Sources

- [Story source](<https://arxiv.org/abs/2611.18586>)

