These tools and metrics are designed to help AI actors develop and use trustworthy AI systems and applications that respect human rights and are fair, transparent, explainable, robust, secure and safe.
Unveiling Covert Toxicity in Multimodal Data via Toxicity Association Graphs: A Graph-Based Metric and Interpretable Detection Framework
This tool provides a graph-based metric and interpretable detection framework to identify covert toxicity that can arise when multimodal signals are combined. It measures toxicity associations across modalities and produces results that can be interpreted to support investigation of safety risks. It is intended for use by practitioners and evaluators aiming to address safety and transparency objectives in multimodal AI systems, and to support accountability and robustness checks.
Auto-discovered on 2026-09-23 by OECD Catalogue Automation
About the metric
You can click on the links to see the associated metrics
Metric type(s):
Objective(s):
Lifecycle stage(s):
Usage rights:
Target users:
Partnership on AI
