Toxicity Detection can be Sensitive to the Conversational Context

TL;DR

This paper introduces a new dataset and task for detecting context-sensitive toxicity in online posts, demonstrating that machine learning models can be trained to identify posts whose toxicity perception depends on conversational context.

Contribution

The paper creates a novel dataset with context-aware toxicity labels and proposes a new task to estimate context sensitivity, improving toxicity detection methods.

Findings

Classifiers can be trained to identify context-sensitive toxicity.

Data augmentation with knowledge distillation enhances detection performance.

Systems can inform moderators about when context is necessary.

Abstract

User posts whose perceived toxicity depends on the conversational context are rare in current toxicity detection datasets. Hence, toxicity detectors trained on existing datasets will also tend to disregard context, making the detection of context-sensitive toxicity harder when it does occur. We construct and publicly release a dataset of 10,000 posts with two kinds of toxicity labels: (i) annotators considered each post with the previous one as context; and (ii) annotators had no additional context. Based on this, we introduce a new task, context sensitivity estimation, which aims to identify posts whose perceived toxicity changes if the context (previous post) is also considered. We then evaluate machine learning systems on this task, showing that classifiers of practical quality can be developed, and we show that data augmentation with knowledge distillation can improve the performance further. Such systems could be used to enhance toxicity detection datasets with more context-dependent posts, or to suggest when moderators should consider the parent posts, which often may be unnecessary and may otherwise introduce significant additional cost.