Source: TechPolicy.Press

TechPolicy.Press explores the human rights implications of content moderation practices powered by large language models (LLMs). The article warns that these AI-driven systems may suppress lawful expression, particularly from marginalized groups, due to opaque filtering criteria and lack of human oversight.

Experts cited in the article call for safeguards such as transparency reports, external audits, appeal mechanisms, and regulatory frameworks aligned with international human rights standards. They argue that without legal accountability, automated moderation may infringe on freedom of speech, access to information, and non-discrimination principles.

The authors stress the urgency of integrating human rights impact assessments into the design and deployment of LLMs used for content filtering. Regulatory action is recommended to prevent misuse and ensure that technology respects democratic values and fundamental freedoms.