# Bias Beneath the Tone: Empirical Characterisation of Tone Bias in LLM-Driven UX Systems

arXiv: https://arxiv.org/abs/2512.19950 · HTML: https://mlsanity.com/paper/2512.19950/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 30 references that are in the map.
- First posted on arXiv 23 Dec 2025.

## Builds on (most cited first, 15 of 15)

- [Bias and Fairness in Large Language Models: A Survey](https://mlsanity.com/paper/2309.00770/) (2023)
- [DialogueRNN: An Attentive RNN for Emotion Detection in Conversations](https://mlsanity.com/paper/1811.00405/) (2018)
- [DialogueGCN: A Graph Convolutional Neural Network for Emotion Recognition in Conversation](https://mlsanity.com/paper/1908.11540/) (2019)
- [Topic-Driven and Knowledge-Aware Transformer for Dialogue Emotion Detection](https://mlsanity.com/paper/2106.01071/) (2021)
- [EmoCaps: Emotion Capsule based Model for Conversational Emotion Recognition](https://mlsanity.com/paper/2203.13504/) (2022)
- [A Transformer-Based Model With Self-Distillation for Multimodal Emotion Recognition in Conversations](https://mlsanity.com/paper/2310.20494/) (2023)
- [BiERU: Bidirectional Emotional Recurrent Unit for Conversational Sentiment Analysis](https://mlsanity.com/paper/2006.00492/) (2020)
- [AI-AI Bias: large language models favor communications generated by large language models](https://mlsanity.com/paper/2407.12856/) (2024)
- [Customizing Emotional Support: How Do Individuals Construct and Interact With LLM-Powered Chatbots](https://mlsanity.com/paper/2504.12943/) (2025)
- [Neural Semantic Encoders](https://mlsanity.com/paper/1607.04315/) (2016)
- [CFN-ESA: A Cross-Modal Fusion Network with Emotion-Shift Awareness for Dialogue Emotion Recognition](https://mlsanity.com/paper/2307.15432/) (2023)
- [Mind Your Tone: Investigating How Prompt Politeness Affects LLM Accuracy (short paper)](https://mlsanity.com/paper/2510.04950/) (2025)
- [ChatGPT Reads Your Tone and Responds Accordingly -- Until It Does Not -- Emotional Framing Induces Bias in LLM Outputs](https://mlsanity.com/paper/2507.21083/) (2025)
- [A Comprehensive Survey of Text Classification Techniques and Their Research Applications: Observational and Experimental Insights](https://mlsanity.com/paper/2401.12982/) (2024)
- [The Impact of Perceived Tone, Age, and Gender on Voice Assistant Persuasiveness in the Context of Product Recommendations](https://mlsanity.com/paper/2405.04791/) (2024)

## Built on by (most cited first, 1 of 1)

- [From 'What' to 'How' and 'Why': Sharing LLM-Generated Retrospective Summaries of Older Adults' Passive Tracking Data with Remote Family Members](https://mlsanity.com/paper/2606.03876/) (2026)

## Abstract (arXiv)

Large Language Models are increasingly used in conversational systems such as digital personal assistants, shaping how people interact with technology through language. While their responses often sound fluent and natural, they can also carry subtle tone biases such as sounding overly polite, cheerful, or cautious even when neutrality is expected. These tendencies can influence how users perceive trust, empathy, and fairness in dialogue. In this study, we explore tone bias as a hidden behavioral trait of large language models. The novelty of this research lies in the integration of controllable large language model based dialogue synthesis with tone classification models, enabling robust and ethical emotion recognition in personal assistant interactions. We created two synthetic dialogue datasets, one generated from neutral prompts and another explicitly guided to produce positive or negative tones. Surprisingly, even the neutral set showed consistent tonal skew, suggesting that bias may stem from the model's underlying conversational style. Using weak supervision through a pretrained DistilBERT model, we labeled tones and trained several classifiers to detect these patterns. Ensemble models achieved macro F1 scores up to 0.92, showing that tone bias is systematic, measurable, and relevant to designing fair and trustworthy conversational AI.