How AI Is Exacerbating Technology-Facilitated Violence Against Women and Girls
- Categories AI, Case Studies
- Date April 28, 2026
How AI Is Exacerbating Technology-Facilitated Violence Against Women and Girls
A case study on artificial intelligence as both a driver of harm and a tool for prevention
98% of deepfakes are pornographic · 99% target women
Analysis and critique by EvalCommunity: This case study presents an independent analysis of the UN Women report “How AI Is Exacerbating Technology-Facilitated Violence against Women and Girls” (2025). It is based on the original publication by UN Women and the Everyday Sexism Project and reflects the perspectives of EvalCommunity on strengths, gaps, and recommendations for M&E practice. EvalCommunity does not claim ownership of the original content. We encourage readers to consult the original source for complete context.
Main reference: UN Women & Everyday Sexism Project (2025). How AI is Exacerbating Technology-Facilitated Violence against Women and Girls. New York: UN Women. Download full report (PDF)
Context & Problem Statement: AI as an Accelerator of Harm
Artificial intelligence is not neutral. Built on human-generated data and deployed without adequate safeguards, AI systems accelerate online violence, deepfake abuse, disinformation, and harassment targeting women and girls worldwide. The UN Women report (2025) documents how AI technologies — from large language models to generative image tools — are being weaponised to perpetrate gender-based violence at scale and speed never before possible.
How AI Magnifies Harm: Five Accelerators
- Scale: Automated generation of thousands of abusive messages, fake profiles, and manipulated images in minutes.
- Speed: Deepfakes and disinformation spread globally before any meaningful response.
- Anonymity: Bots and AI-generated personas make perpetrators harder to identify and prosecute.
- Affordability: Low-cost or free tools for deepfakes, voice cloning, and impersonation (as low as $0.34 per deepfake).
- Realism: Highly convincing synthetic media enables scams, political manipulation, and reputational harm at scale.
Forms of Violence Amplified by AI
AI bots amplify malicious content, automate trolling, and reinforce stereotypes via biased training data. Over 65% of women witnessed this form of abuse.
AI-enhanced catfishing, voice cloning, and account hacking: 63% of women experienced or witnessed impersonation-based abuse.
98% of deepfake videos are pornographic; 99% target women. Deepfakes are used as political violence to silence women in public life.
AI scrapes personal data (addresses, contacts) to assemble profiles for stalking and offline intimidation. 50%+ witnessed threat tactics.
Structural Inequality Behind the Crisis
countries reference gender in AI strategies
include gender-responsive provisions
of AI professionals are women
women internet access in low-income countries
When women are absent from design, bias and harm are baked into systems. The absence of gender perspectives in national AI strategies risks normalizing discrimination at scale.
AI for Good: Emerging Prevention Tools
24/7 digital ally for domestic violence survivors: legal options, evidence storage, referrals.
AI chatbot providing legal advice to survivors of digital sexual violence (Mexico/Ecuador).
NLP-powered dashboard detects online harassment, co-designed with 230+ young women.
WhatsApp tool for West/Central Africa, supports TFGBV survivors in multiple local languages.
The Way Forward: 5 Pillars of Gender-Responsive AI
Safety by design
Rapid response
Women shaping AI
Clear liability
Closing digital divides
Legal & Policy Levers for Change
- Criminalize non-consensual sexual deepfakes and impersonation (EU AI Act transparency provisions as model).
- Mandatory Gender Impact Assessments for all high-risk AI systems before deployment.
- Platform accountability: proactive detection of coordinated harassment, cross-platform cooperation, transparency reports on gender-based harms.
- Civil remedies for survivors and ringfenced statutory funding for support services.
- Intersectional protections: marginalized women face higher risks and need tailored safeguards.
Implications for M&E Practice in AI & Digital Safety
Evaluators must track both harm reduction (removals, blocks) and unintended consequences (false positives, censorship).
Without sex-disaggregated and intersectional data, AI violence remains invisible in official statistics.
Dashboards on deepfake reports, platform response times, and survivor feedback loops should be standard M&E tools.
Evaluations must prioritise survivor safety, do-no-harm principles, and informed consent.
M&E Challenge: Apply the Evidence
1. Only 24 of 138 countries reference gender in national AI strategies. What is the primary M&E risk?
2. Deepfake abuse: 98% of deepfake videos are pornographic, 99% target women. Which policy lever is most effective?
3. Women are only 30% of AI professionals. What indicator best tracks inclusive AI governance?
Key takeaway for M&E and development practitioners
The future of AI is not predetermined. It is a policy choice. Without gender-responsive safeguards, AI will deepen inequality and normalize violence. Evaluators must track representation metrics, deepfake legislation, platform duty-of-care compliance, and survivor-centered justice indicators. The most urgent shift: from reactive content removal to structural prevention and gender-by-design.
References & Sources
- UN Women & Everyday Sexism Project (2025). How AI is Exacerbating Technology-Facilitated Violence against Women and Girls. New York: UN Women.
- UN Women AI School: unwomen.org.au
- EvalCommunity Academy case studies: StopNCII, Kemi, Sophia, OlimpiA, AI Ally.
Disclaimer: Analysis and critique by EvalCommunity. This case study is an independent analysis for educational and policy guidance purposes. EvalCommunity does not claim ownership of the original UN Women content. Readers are encouraged to consult the primary source for complete methodology and context.
Published by EvalCommunity Academy. Based on UN Women research (2025) under CC BY-NC-ND 3.0 IGO. For professional M&E training on AI ethics and gender-based violence prevention, visit EvalCommunity Academy.
The courses and articles are developed by a team of experienced evaluators, collaborators, authors, and software developers, guided by Fation Luli. EvalCommunity Academy combines practical expertise in Monitoring & Evaluation and International Development with the latest advances in AI to create high-quality, accessible, and practical learning experiences for professionals worldwide.