Regulating Online Hate Speech through the Prism of Human Rights Law: The Potential of Localised Content Moderation

Abstract

Abstract This article explores whether international human rights standards can provide insights into the following questions: who can and should define what constitutes ‘hate speech’ for online moderation, and how can hate speech be detected and moderated on social media platforms? Regarding who should moderate hateful content, the article underscores the responsibility of social media companies, in reference to international human rights law and principles on business and human rights. This paper advocates for a multistakeholder approach where companies work in partnership with and under the monitoring of state actors, civil society and other relevant stakeholders. Given the complexity of what constitutes hate speech, the article proposes the localisation of terms of service and guidelines of social media companies. This approach, in line with human rights standards, enables the meaningful involvement of local experts and civil society groups in formulating, implementing and monitoring online community rules. Addressing the question of how social media companies can detect and moderate hate speech, the article argues that a globally standardised approach reliant on AI content moderation is limited in detecting contextual nuances of hate speech. Drawing from international human rights tools like the Rabat Plan of Action, the article suggests that social media companies should consider diverse local contexts for better interpretation and effective detection, with qualified human moderators with local knowledge and oversight boards with local expertise. By taking a human-centered, localised approach to content moderation and collaborating with relevant stakeholders, businesses can contribute to creating a digital space that upholds fundamental rights, prevents harm, and encourages inclusivity and diversity.