On the Role of Speech Data in Reducing Toxicity Detection Bias
This work addresses bias in toxicity detection systems for users of social media or content moderation, but it is incremental as it builds on existing datasets and methods.
The study tackled bias in toxicity detection by comparing speech- and text-based systems, finding that using speech data during inference reduces bias against demographic group mentions, especially for ambiguous cases, with a focus on improving classifiers over transcription pipelines.
Text toxicity detection systems exhibit significant biases, producing disproportionate rates of false positives on samples mentioning demographic groups. But what about toxicity detection in speech? To investigate the extent to which text-based biases are mitigated by speech-based systems, we produce a set of high-quality group annotations for the multilingual MuTox dataset, and then leverage these annotations to systematically compare speech- and text-based toxicity classifiers. Our findings indicate that access to speech data during inference supports reduced bias against group mentions, particularly for ambiguous and disagreement-inducing samples. Our results also suggest that improving classifiers, rather than transcription pipelines, is more helpful for reducing group bias. We publicly release our annotations and provide recommendations for future toxicity dataset construction.