This project involves developing an AI-powered chatbot to combat misinformation affecting migrant children and youth in the United States. Through a B2B2C model, the tool targets intermediary organizations (educators, social workers, journalists, NGOs, and parents) who act as validators and distributors of verified information rather than directly reaching youth. Using RAG architecture and natural language processing in Spanish, the chatbot integrates Factchequeado's curated database with governmental and civil organization sources to provide accurate, verified information on migration-related topics. The development will be open-source and scalable across partner organization websites, prioritizing quality over quantity in content curation and simple solutions that respond to the "triple distrust" and "ideological vulnerability" patterns identified in target populations.

Empower migrant children and youth in the United States to access verified, trustworthy information about migration through an open-source, AI-powered platform that combats misinformation and strengthens digital trust.

**Primary Goal: **Develop and deploy an AI-powered chatbot system that addresses the critical information crisis facing migrant children, youth, and their families by providing accurate, verified migration-related information through trusted intermediary organizations.
**Target Impact: **Reduce the harmful effects of misinformation on vulnerable migrant populations by creating a scalable, open-source solution that connects verified information sources to those who need them most, while building digital literacy and critical thinking capabilities within communities.
**Strategic Approach: **Leverage a B2B2C model where credible intermediary organizations (educators, social workers, journalists, NGOs, parents) serve as validators and distributors of information, addressing the "triple distrust" pattern (distrust in media, immigration systems, and institutions) by working through already-trusted community touchpoints.
Innovation Focus: Create cutting-edge open-source technology at the intersection of artificial intelligence and data trust that can be scaled globally to benefit children in UNICEF's 190+ country network, contributing to the broader Digital Public Goods ecosystem.

The core problem identified in the project is that migrant children and youth in the United States face a critical information crisis characterized by widespread misinformation and significant barriers to accessing accurate, verified information about migration-related topics. This problem manifests through a "triple distrust" pattern—distrust in media, the immigration system, and institutional organizations—combined with "ideological vulnerability" where communities place greater trust in individuals and influencers (often on platforms like TikTok and WhatsApp) than in official sources, making them susceptible to misinformation. The information ecosystem is further complicated by rapid viral spread of false information, constant changes in immigration policies, lack of resources to verify information, language barriers (most information is in English), state-specific laws requiring localized context, fear within communities, and prevalent scams targeting vulnerable populations. Children and youth specifically struggle to access critical information about legal status and rights, how government systems work, health and mental health services, educational access, fraud prevention, and how to protect parents in mixed-status families. This information gap is exacerbated by the lack of cross-border databases for transnational migration issues and few organizations serving migrant communities in certain states, creating an urgent need for a trusted system that delivers verified information through credible intermediary organizations.

The proposed solution is an AI-powered web chatbot that serves as a "smart search widget" with natural language query capability in Spanish, built on RAG (Retrieval Augmented Generation) architecture. The chatbot integrates Factchequeado's manually curated database of verified migration information with official government sources and civil organization content, allowing users to ask questions in natural language and receive accurate, source-grounded responses. The system is designed for easy integration into partner organization websites through a simple embeddable widget, prioritizing scalability and open-source development. Key technical features include a verified migration information repository with official policy documents, misinformation detection and fact-checking classification systems, interactive visualizations to explain migration processes, and multimedia content integration including videos, infographics, and educational materials. The platform includes a database administration interface for managing documents, lexicons, and other data components, with stress testing to ensure the system appropriately responds "I don't know" when queries fall outside its knowledge base. The solution is deliberately simple rather than complex, focusing on a single web-based chatbot rather than multiple platforms or course systems, and targets intermediary organizations rather than children directly to avoid privacy concerns while maximizing reach through trusted validators who already have credibility with migrant communities.