Much of the online violence against women in West and Central Africa happens on platforms owned by global technology companies. RFLD’s report Online violence against women in West and Central Africa 2026 devotes a chapter to how these platforms respond, and to the new risks that artificial intelligence brings.
A language gap in moderation
Automated content moderation covers fewer than 20 of Africa’s roughly 2,000 languages. Only 42 African languages appear in any meaningful way in the major language models reviewed in a 2025 study. Four have consistent coverage: Amharic, Swahili, Afrikaans and Malagasy. Hausa and Yorùbá are among those with some coverage.
No West or Central African language has been the subject of a dedicated moderation study. The Center for Democracy and Technology’s series on content moderation in the Global South studied Kiswahili and Maghrebi Arabic in Africa, but none of the region’s major languages, such as Hausa, Wolof, Lingala, Bambara, Fon or Ewe. The report names this as a research gap.
Abuse written in Wolof or Fon, or in local forms of French, is therefore less likely to be detected automatically.
Little country-level data
Platforms rarely publish enforcement data for the region. One exception comes from Senegal, where TikTok removed more than 2.5 million videos and interrupted more than 16,000 live sessions for policy violations between early 2024 and mid-2025. By comparison, 450,000 videos were reported removed in Kenya in January to March 2025.
User experience is also a concern. In Nigeria, Gatefield’s research found that only 24% of users find X responsive to complaints about online harm. In the same research, X accounted for 34% of documented online harms in Nigeria, ahead of Facebook at 29%.
Sexualised deepfakes
Artificial intelligence has made it easy to create fake sexual images. UN Women, citing Sensity.ai, reports that 90 to 95% of deepfakes online are sexual images of women. UN Women’s 2025 Tipping point survey found that 24% of respondents, women defenders and journalists among them, had faced AI-assisted abuse.
In Nigeria, AI-generated sexual images of well-known women artists circulated in 2025 without visible platform action. Gatefield projects that, without intervention, tens of millions of Nigerian women and girls could face AI-facilitated abuse each year by 2030. The report is careful to describe this as a modelled projection, not a measured figure.
Tools that work
Some tools already help. Hash-matching services let an adult create a digital fingerprint of an intimate image on her own device, so the image itself is never uploaded. Participating platforms, including Meta and TikTok, then block matching uploads. StopNCII.org is the best-known example.
Reporting directly to platforms can also work. In the 2020 Pollicy study, 75% of women in Senegal who reported abuse to a platform saw the content removed.
Where accountability lies
African law says little about platforms. ARTICLE 19 has noted that the AU Convention on Ending Violence against Women and Girls places accountability on states, with no obligations on platforms. The main African soft-law source on intermediaries is Principle 39 of the African Commission’s 2019 Declaration on Freedom of Expression, which asks them to mainstream human rights safeguards and to be transparent about content removal requests.
The Budapest Convention on Cybercrime, to which eight West and Central African states are parties, offers cooperation channels for obtaining evidence from foreign platforms. Resolution 522 asks states to facilitate cooperation between law enforcement and service providers, in line with data protection standards.
What the report asks of technology companies
The report makes four recommendations to technology companies:
- invest in content moderation in the languages of the region and in local forms of French, and publish country-level data on reports and how they are handled;
- create priority reporting channels for women in public life, especially during elections;
- participate in hash-matching tools and extend them to messaging services where technically possible;
- label and remove sexualised deepfakes promptly, and support survivors in removing content across platforms.
Messaging services matter in particular. In Benin, WhatsApp accounted for 21% of cases in a 2025 WILDAF-AO report. Abuse that circulates in private groups is hard to reach with tools designed for public posts.
Further reading
The report summary sets out all the recommendations.
About the report
This article draws on RFLD’s report Online violence against women in West and Central Africa 2026, published on 4 October 2026. Read the summary, download the English edition or the French edition.
