arXiv:2502.08266v3 Announce Type: replace-cross Abstract: Hate speech detection is a crucial task, especially on social media where harmful content can spread quickly. Collecting social media content (tweets etc.) to train machine learning models is easy, but detecting and categorizing hate speech can be difficult due to the inherently subjective nature. This subjectivity leads to frequent disagreement among annotators, particularly for subtle or borderline content. Traditional approaches either discard non-consensus samples or force a ''gold standard'' through expert adjudication, ignoring va
The proliferation of social media platforms and the increasing reliance on AI for content moderation make robust hate speech classification a pressing need.
Improving hate speech detection directly impacts online safety, platform governance, and the ethical deployment of AI in sensitive social contexts.
This research highlights the inherent subjectivity in hate speech annotation and proposes methods to explicitly model annotator disagreement, moving beyond simplistic 'gold standard' approaches.
- · Social media platforms
- · AI ethics researchers
- · Trust and Safety teams
- · Content creators using subtle hate speech
- · Traditional fixed-label classification models
More nuanced and effective hate speech detection models are developed and deployed across platforms.
Social media content moderation policies evolve to integrate and reflect the complexities of annotator disagreement, leading to fairer and more consistent application.
Enhanced understanding of hate speech subjectivity informs broader discussions on online discourse, free speech, and the role of AI in shaping public opinion.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG