Turning AI risk signals into actionable intelligence
We continuously monitor frontier AI capabilities, incidents, and policy developments to help policymakers understand emerging AI risks before they become crises.
AIRE is an intelligence platform for frontier AI risk.
We aggregate, interpret, and organize the evidence on dangerous AI capabilities into a single, continuously updated resource.
AIRE synthesizes frontier model evaluations, real-world incidents, and governance developments into risk intelligence that is maintained over time — not published once and left to age.
The evidence is scattered.
Today, information about frontier AI risks is fragmented across research papers, incident reports, and policy documents. Decision-makers rarely have the time to synthesize this evidence.
AIRE continuously aggregates, interprets, and organizes this information into actionable intelligence.
From raw evidence to intelligence
Collect
We compile evidence from system cards, third-party evaluations, incident reporting, regulatory documents, and other open data.
Structure
We tag every entry against taxonomies and schemas across all databases, providing a common language for analysis.
Assess
We mine the accumulated evidence to track risk evolution, identify emerging patterns, and answer questions no single source can.
Disseminate
We publish assessments, dashboards, and briefings that give decision-makers a map of the AI risk landscape.
A living platform, not a report.
Each repository feeds a view built on top of it. Hover a product to see what is inside it right now.
Model Evaluations Database
A continuously updated repository of AI risk evaluations, with the key findings and risk levels extracted from each.
Frontier AI Risk Assessments
Expert synthesis of frontier model capabilities across our four risk domains, scored against defined risk thresholds.
Four risk domains
We hold deep focus on the four vectors of AI-enabled harm with the greatest potential for large-scale impact.
The people behind AIRE
A multidisciplinary group spanning analysis, operations, and engineering.
Guillem Bas
Guillem leads the project's strategic direction and coordinates analysis, engineering, and operations. He has previously worked on EU AI policy, contributing to the AI Act and the GPAI Code of Practice, and has published research on international AI governance.
Husanjot Chahal
Husanjot researches the intersection of AI, international security, and technology governance. She has previously worked at OpenAI, Georgetown's CSET, and the World Bank, and her research has been featured by NATO, the European Parliament, Politico, and Fortune.
Andrea Castillo
Andrea runs finance and administration, from cash flow and management reporting to audits. She brings sixteen years of experience across construction, banking, mining, and the non-profit sector.
Tim Sankara
Tim builds and maintains the platform, its databases, and the pipelines that keep them current. He is a full-stack engineer with over nine years of experience in data-driven applications, and has completed advanced training in AI safety and alignment.
Michelle Bruno
Michelle leads our work on biological risk, translating developments in the life sciences into biosecurity analysis. She is a genomics biotechnologist and a member of the Mexican Biosecurity Association (AMEXBIO) and the iGEM Safety Committee.
Stay ahead of AI risk.
Explore continuously updated intelligence on frontier model evaluations, real-world incidents, and the policies shaping what comes next.