Lune

EMNLP Findings2025

Low-Resource Languages LLM Disinformation is Within Reach: The Case of Walliserdeutsch

Andrei Kucharavy, Sherine Seppey, Cyril Vallez, Dimitri Percia David, Ljiljana Dolamic

Abstract

LLM-augmented online disinformation is of particular concern for low-resource languages, given their prior limited exposure to it. While current LLMs lack fluidity in such languages, their multilingual and emerging capabilities can potentially still be leveraged. In this paper, we investigate whether a moderately sophisticated attacker can leverage such capabilities and perform an impersonation attack in the Walliserdeutsch dialect, a low-resource (100k speakers) Swiss German Highest Allemanic dialect that is generally non-intelligible to both Standard German and other Swiss German dialects speakers and presents considerable within-dialect variability.

We show that while a standard few-shot learning prompting of SotA LLMs, even by native Walliserdeutsch speakers, yields easily humandetectable texts, an expert attacker performing a PEFT on a small SotA LLM is partially able to perform such an impersonation with minimal resources, even if the fine-tuned LLM does not advertise any capabilities in Germanic languages. With Walliserdeutsch presenting many features of low-resource languages and dialects, our results suggest that LLM-augmented disinformation is within reach for low-resource languages, highlighting the urgency of LLM detectability research in low-resource languages.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Language Model Tokenizers Introduce Unfairness Between Languages cites MEGA: Multilingual Evaluation of Generative AI
  • Not All Languages Are Created Equal in LLMs: Improving Multilingual Capability by Cross-Lingual-Thought Prompting cites MEGA: Multilingual Evaluation of Generative AI
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on9

  1. LoRA: Low-Rank Adaptation of Large Language ModelsEdward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu et al.ICLR 2022 · 18,833 citations
  2. Language Model Tokenizers Introduce Unfairness Between LanguagesAleksandar Petrov, Emanuele La Malfa, Philip H. S. Torr, Adel BibiNeurIPS 2023 · 301 citations
  3. MEGA: Multilingual Evaluation of Generative AIKabir Ahuja, Harshita Diddee, Rishav Hada, Millicent Ochieng et al.EMNLP 2023 · 91 citations
  4. Not All Languages Are Created Equal in LLMs: Improving Multilingual Capability by Cross-Lingual-Thought PromptingHaoyang Huang, Tianyi Tang, Dongdong Zhang, Xin Zhao et al.EMNLP Findings 2023 · 76 citations
  5. MasakhaNER 2.0: Africa-centric Transfer Learning for Named Entity RecognitionDavid Ifeoluwa Adelani, Graham Neubig, Sebastian Ruder, Shruti Rijhwani et al.EMNLP 2022 · 46 citations

Related papers

  1. Dialect Transfer for Swiss German Speech TranslationClaudio Paonessa, Yanick Schraner, Jan Deriu, Manuela Hürlimann et al.EMNLP Findings 2023 · 3 citations
  2. Dialectal Toxicity Detection: Evaluating LLM-as-a-Judge Consistency Across Language VarietiesFahim Faisal, Md Mushfiqur Rahman, Antonios AnastasopoulosEMNLP Findings 2025 · 3 citations
  3. It's All About In-Context Learning! Teaching Extremely Low-Resource Languages to LLMsYue Li, Zhixue Zhao, Carolina ScartonEMNLP 2025 · 1 citation
  4. RAFT: Realistic Attacks to Fool Text DetectorsJames Wang, Ran Li, Junfeng Yang, Chengzhi MaoEMNLP 2024 · 2 citations
  5. AustroTox: A Dataset for Target-Based Austrian German Offensive Language DetectionPia Pachinger, Janis Goldzycher, Anna Maria Planitzer, Wojciech Kusa et al.ACL Findings 2024 · 2 citations
Dusk over the sea between two cliffs drawn in fine vertical lines