Run the research.

A multi-party AI negotiation table where each delegation argues from its own confidential 'privileged instructions.' Five crises — Arctic, Central Asia, Cyprus, South China Sea, and Iran–US Strait of Hormuz — each with a live-negotiation replay, an interactive analysis dashboard (parties' strategy, proposals, detector hits), and Monte Carlo runs over randomized conditions.

Three civilian-evacuation risk tools behind one backend and design language: a crisis map of the INFORM Severity Index with live news and conflict timelines, a seven-dimension scenario risk model, and an endangerment-and-feasibility assessment.

A real-time index measuring the quality, accessibility, and equity of evacuation information during crises — synthesizing agency, social, and on-the-ground signals to surface where critical information is missing.

An interactive simulator modeling how routing and assignment choices shape who reaches safety, with shareable scenario state and a built-in explainer walkthrough.

A corridor-severity model reconstructing the 2022 Mariupol evacuation, scoring the risk and viability of humanitarian corridors under siege conditions.

An agent-based model of community evacuation behavior — family clusters, information-seeking, neighbor influence, and elder and child delays across an UNAWARE → SEEKING → MILLING → EVACUATING → DONE lifecycle.

A digital provenance passport for cultural artifacts, pairing verifiable credentials with collaborative cataloguing to support repatriation, attribution, and ethical stewardship.

An evacuation-modeling prototype exploring population movement and decision dynamics under crisis constraints.

A decision-support framework estimating the human and financial cost of civilian evacuation in armed conflict, scoring scenarios across seven risk dimensions and comparing evacuating against staying.

A composite index anticipating civilian risk during evacuation operations, scoring weighted categories — hazards, infrastructure, population vulnerability, information environment, and weather — with Monte Carlo sensitivity analysis, ACLED conflict-event lookup, and source-credibility tagging on every indicator, so evacuation-risk reasoning stays auditable for IHL compliance review. Built by Alana Robertson in the Summer 2026 cohort, on Teresa Cantero's doctoral research into AI and the protection and evacuation of civilians under IHL.

An x402-native agent that traces artwork provenance, flags looting and repatriation risk, and issues a signed, tamper-evident Passport for each object. This live demo runs on mock data — catalog dashboards, passport issue/verify, and a replayed agent trace.

An interactive index mapping the structural conditions that enable forced labor across 184 countries on a 0–1 risk scale — with a choropleth world map, national and sub-national layers, rankings, and sources. Built by Amanda Lindsey in the Fall 2025 cohort.

A survey and toolkit for how AI helps researchers move from a broad interest to a well-formed research question — finding gaps, generating candidate questions, summarizing the state of the art, and flagging contradictions, with guardrails against hallucinated citations. Ships with a reusable Copilot 'Researcher' prompt and a journal-credibility rubric. Built by the Fall 2025 cohort.

Human-condition storytelling with culturally grounded digital-human avatars — a set of short pieces produced with generative media in the Spring 2025 cohort.