Global Partnership on AI-Enabled Manipulation & Scams
Developing behavioral AI methods to detect and prevent AI-enabled manipulation, fraud, and psychological exploitation across digital platforms.
Partners
Shisa.AI
Shisa.AI
Status
Status
Ongoing
Ongoing
Focus
Focus
AI-Enabled Manipulation
AI-Enabled Manipulation
Overview
The Global Partnership on AI-Enabled Manipulation & Scams is an international initiative dedicated to understanding, detecting, and preventing AI-enabled psychological manipulation and online fraud.
As generative AI becomes increasingly capable of conducting persuasive, personalized conversations, new forms of fraud, social engineering, and digital exploitation are emerging at an unprecedented scale. This project brings together behavioral science, artificial intelligence, and cross-sector collaboration to develop practical methods for identifying manipulation before significant harm occurs.
Selected by the United Nations AI Dialogue Partnerships Hub, the initiative aims to support safer digital ecosystems through open research, evaluation frameworks, and international partnerships.
Research Approach
Our research combines behavioral science with advanced AI evaluation to better understand how manipulation unfolds in real-world digital conversations.
The project analyzes large-scale multilingual scam conversations to identify behavioral patterns, cognitive biases, persuasion strategies, and escalation pathways used by sophisticated fraud networks. These insights are used to develop behavioral evaluation benchmarks, specialized detection models, and practical guidance for governments, technology companies, financial institutions, and civil society organizations.
Rather than focusing solely on language models themselves, the project examines how AI influences human judgment, trust, and decision-making throughout extended interactions.
Expected Inpact
The Global Partnership seeks to establish an international foundation for protecting people from AI-enabled manipulation while supporting responsible AI innovation.
Expected outcomes include open behavioral evaluation frameworks, multilingual datasets, practical guidance for AI developers and policymakers, collaborative workshops across multiple countries, and stronger partnerships between academia, industry, governments, and international organizations.
By placing human behavior at the center of AI safety and governance, the project aims to help organizations detect emerging risks earlier, design more trustworthy AI systems, and strengthen resilience against rapidly evolving forms of digital manipulation and fraud.
RESEARCH PROJECTS
Research at the Intersection of Human Behavior and AI.
Global Partnership on AI-Enabled Manipulation & Scams
Developing behavioral AI methods to detect and prevent AI-enabled manipulation, fraud, and psychological exploitation across digital platforms.
Partners
Shisa.AI
Shisa.AI
Status
Status
Ongoing
Ongoing
Focus
Focus
AI-Enabled Manipulation
AI-Enabled Manipulation
Overview
The Global Partnership on AI-Enabled Manipulation & Scams is an international initiative dedicated to understanding, detecting, and preventing AI-enabled psychological manipulation and online fraud.
As generative AI becomes increasingly capable of conducting persuasive, personalized conversations, new forms of fraud, social engineering, and digital exploitation are emerging at an unprecedented scale. This project brings together behavioral science, artificial intelligence, and cross-sector collaboration to develop practical methods for identifying manipulation before significant harm occurs.
Selected by the United Nations AI Dialogue Partnerships Hub, the initiative aims to support safer digital ecosystems through open research, evaluation frameworks, and international partnerships.
Research Approach
Our research combines behavioral science with advanced AI evaluation to better understand how manipulation unfolds in real-world digital conversations.
The project analyzes large-scale multilingual scam conversations to identify behavioral patterns, cognitive biases, persuasion strategies, and escalation pathways used by sophisticated fraud networks. These insights are used to develop behavioral evaluation benchmarks, specialized detection models, and practical guidance for governments, technology companies, financial institutions, and civil society organizations.
Rather than focusing solely on language models themselves, the project examines how AI influences human judgment, trust, and decision-making throughout extended interactions.
Expected Inpact
The Global Partnership seeks to establish an international foundation for protecting people from AI-enabled manipulation while supporting responsible AI innovation.
Expected outcomes include open behavioral evaluation frameworks, multilingual datasets, practical guidance for AI developers and policymakers, collaborative workshops across multiple countries, and stronger partnerships between academia, industry, governments, and international organizations.
By placing human behavior at the center of AI safety and governance, the project aims to help organizations detect emerging risks earlier, design more trustworthy AI systems, and strengthen resilience against rapidly evolving forms of digital manipulation and fraud.
RESEARCH PROJECTS