What is the challenge

Conflict‑related sexual violence (CRSV) is one of the most devastating, underreported, and poorly monitored violations in crisis settings. Survivors, including women, girls, LGBTQ+ individuals, prisoners of war, and other marginalised groups, face significant barriers to accessing justice, protection, and essential services. Insecurity, stigma, weak legal systems, and widespread impunity further limit reporting and accountability.

Across contexts such as the Democratic Republic of Congo, South Sudan, Myanmar, and Ukraine, UN Human Rights has documented alarming patterns of CRSV, with over 70 percent of listed perpetrators remaining non‑compliant for five years or more. Gender-based violence in conflict, post-conflict and humanitarian settings is rising. Full reporting of every case is impossible, particularly when humanitarian, human rights and NGO staff are attacked in conflict zones and prevented from accessing them.

Despite growing demand for early warning and anticipatory action, CRSV data remains fragmented across open sources, partner datasets, civic space indicators, and information on arms and ammunition flows. Current monitoring systems often fail to capture escalation early enough to shape humanitarian response or inform survivor‑centred protection planning. The UN Secretary‑General has called for strengthened monitoring, survivor‑centred approaches, and the use of innovative technologies to support documentation, analysis, and accountability. At the same time, cuts to humanitarian and development aid by international donors are impacting prevention programmes.

This project responds directly to these gaps by by enabling faster, rights‑respecting detection of CRSV risks and helping humanitarian, development, and peace actors act before violence escalates, and to provide an evidence base to hold perpetrators accountable.

What is innovative about the project

Violet.AI scales a proven, AI‑enabled, privacy‑preserving system developed by the Office of the United Nations High Commissioner for Human Rights (OHCHR) to automate and enhance the detection, analysis, and early warning of CRSV. The system consolidates CRSV‑related information from open sources and trusted partners using machine learning and natural language processing, builds high‑quality canonical datasets, and generates risk scores and trend analysis. It also incorporates synthetic data generation and differential privacy to preserve confidentiality and prevent reidentification while enabling safe insights from sensitive datasets.

A major innovation is its integration of CRSV data with complementary datasets, including arms and ammunition flows from UNIDIR, civic space indicators from OHCHR’s internal Civic Space Pulse platform (such as internet shutdowns and protest repression), and verified incident‑level data from Insecurity Insight. This multi‑layered risk modelling allows Violet.AI to identify enabling conditions and direct threats to communities in real time.

Insights are delivered through a secure dashboard and API with differential access levels, designed for field use by women‑led protection networks, humanitarian responders, and UN early warning systems. The system aligns with JIAF, protection cluster frameworks, and OHCHR’s Human Rights‑Based Approach to Data, ensuring that innovation reinforces, rather than jeopardises, the rights of survivors and affected communities.

What are the expected outcomes

During the project period, Violet.AI will:

• Scale the current pilot from 8 conflict‑affected countries to new geographies with high CRSV risk.
• Integrate new data pipelines, including arms flows, insights on civic space deterioration, and attacks on human rights defenders.
• Deploy secure CRSV dashboards and risk‑scoring tools for humanitarian, development, peace and human rights actors.
• Support anticipatory action through improved early warning and near real‑time insights for protection coordination.
• Strengthen local ownership, including through co‑design workshops with women’s rights groups and survivor networks.

In the longer term, the project aims to:

• Institutionalise CRSV early warning capacity across the UN system.
• Build a sustainable, rights‑respecting protection data infrastructure across 20+ countries.
• Provide a global public good aligned with the Women, Peace and Security agenda.
• Improve compliance with arms control and disarmament frameworks through better data on weapons‑enabled violence.
• Reduce the incidence and severity of CRSV by enabling earlier, more targeted, and coordinated protection interventions.

Who are the project partners

The project is led by OHCHR – Innovation and Analytics Hub, which serves as system developer and provides overall quality assurance, human rights verification, and deployment across UN Human Rights operations.

UN Institute for Disarmament Research (UNIDIR) provides arms and ammunition flow data; integrates weapons‑related insights into CRSV early warning; supports research, modelling, and dashboard development. 

Insecurity Insight supplies high‑frequency incident data on CRSV and related protection threats; supports schema harmonization and trend analysis.

Technology and trusted local partners and women‑led organizations engaged in building digital infrastructure, and in participatory design, validation, and contextualization of protection indicators; support survivor‑centred usage and community adoption.