"Meta’s Oversight Board Issues Unanimous Decision on Threats Against Human Rights Defenders", 28 May 2025

...The Meta Oversight Board announced yesterday its unanimous decision to overturn the company’s initial determination to keep online a post attacking a prominent human rights defender in Peru. The ruling sets a significant precedent in the protection of those who defend fundamental rights amid increasing repression, stigmatization, and the closure of civic space.

The case involved a July 2024 post by a member of the extremist group La Resistencia, known for spreading disinformation, harassment, and incitement against journalists, activists, and civil society organizations in Peru. The post featured a manipulated image—likely generated with artificial intelligence—of a well-known Peruvian human rights defender, depicting her with a bloodied face, accompanied by text discrediting the work of NGOs and linking them to violent protests.

At the Center for Justice and International Law (CEJIL), we reported the post to Meta, arguing that it constituted a veiled threat in a context where online attacks can escalate into physical violence, especially in light of the Peruvian state’s failure to prevent such harms....After Meta chose to keep the post online, we escalated the case to the Oversight Board—an independent body that reviews moderation decisions...

The Board urged Meta to urgently improve its internal processes to better prevent harm in sensitive contexts. It also concluded that a proper interpretation of Meta’s content policies must align with the company’s human rights responsibilities, particularly with the right to freedom of expression as enshrined in Article 19 of the International Covenant on Civil and Political Rights (ICCPR).

The decision includes specific recommendations: Meta should update its policy to explicitly prohibit veiled threats—whether written, visual, or verbal—and conduct annual assessments on this type of content, with special attention to threats against human rights defenders and moderation errors concerning political speech...