About UNU
The United Nations University (UNU) is a global think tank and postgraduate teaching organisation headquartered in Japan. The mission of the UN University is to contribute, through collaborative research and education, to efforts to resolve the pressing global problems of human survival, development and welfare that are the concern of the United Nations, its Peoples and Member States.Activities by UNU
UN AI Co-Policymaker
Global IndonesiaUNU Macau has developed a multi agentic AI tool, rooted in UN frameworks, that advises decision-makers at global, national, and local levels. Named the Co-Policymaker, it engages in interactive dialogues with human users to provide diverse policy options and forecast their impacts based on development economics indicators and ethical frameworks.
Activity Type AI Tools/SolutionsGendered implications of Artificial Intelligence on the implementation of the Women, Peace and Security agenda in Southeast Asia
Africa, Asia and the Pacific Democratic Republic of the CongoKenyaThis research examined the opportunities and risks of AI from a WPS lens in Southeast Asia, with a focus on four types of gender biases in AI which will need to be addressed before the region can fully benefit from new technological developments; discrimination, stereotyping, exclusion, and insecurity. The research included interviews with relevant stakeholders to understand the relationship between AI and WPS as well as a social media analysis of women civil society organisations in the region. These findings were used to inform an e-learning module on AI and the WPS agenda for women civil society in Southeast Asia and two critical recommendations are made; mitigating the risks of AI systems to advancing the WPS agenda and fostering the development of AI tools built explicitly to support gender-responsive peace in line with WPS commitments.
Activity Type Trainings/WorkshopsResearch/Reports/AssessmentsPolicy/Regulatory GuidanceVendor Risk Analysis
Global Japan
The Vendor Risk Analysis (VRA) Workflow is an AI-assisted system designed to produce accurate, audit-ready vendor risk reports through a structured, multi-stage process. It combines AI-driven report generation with two complementary verification methods to ensure reliability and compliance:
Link Analysis validates that every URL cited in the report is accessible and accurately supports the referenced statement, ensuring citation integrity and source trustworthiness.
Claim Analysis independently fact-checks each claim in the report, whether or not it has a cited link, by segmenting content into semantically complete units and verifying them against external sources, ensuring factual correctness.
Multi-agent Multi-model Deep Research System
Global Japan
A multi-agent, multi-model system designed to automate the end-to-end process of deep research and report generation. The system addresses the traditionally labor-intensive and time-consuming nature of knowledge synthesis by orchestrating a sequential, human-in-the-loop workflow.
The architecture comprises a Planning Agent, which deconstructs a high-level user topic into a structured, editable research plan, and a Research Agent, which executes this plan by performing deep web research for each subtopic.
It addresses the limitations of traditional research workflows, which often suffer from algorithmic bias and informational silos, by leveraging multi-engine synthesis and an AI Judge to produce high-fidelity insights. The system ensures objectivity by comparing outputs from different search engines and consolidating them into a single, superior report.
Key Features:
Automated Planning: Breaks down topics into objectives and subtopics.
Deep Web Research: Executes multi-engine searches for broad coverage.
Cited Synthesis: Produces structured reports with references.
AI Judging: Compares multiple drafts and merges the best insights.
Automated Scoring: Quantifies report quality based on coverage and depth.
Iteration Tools: Allows re-synthesis without re-running full research.
Value Proposition:
Delivers comprehensive, unbiased, and actionable research outputs that surpass single-source approaches, making it ideal for academic, policy, and enterprise use cases.
Adaptability Beyond Research:
This architecture can be repurposed for other domains:
Policy Analysis: Aggregate and synthesize legislative data.
Market Intelligence: Compare insights from multiple industry sources.
Risk Assessment: Merge reports from different compliance engines.