About the AI Risk Navigator

AIRI has built some of the most comprehensive datasets on AI risk. But until now, these datasets existed in isolation. A researcher who wanted to know whether governance attention matched expert concern, or whether the risks most frequently studied in the literature were the same ones causing real-world harm, would have had to manually cross-reference across separate Airtable interfaces and published papers.

The Navigator was built to close that gap. By using the risk domain taxonomy as a central navigation point and providing easy-to-use search interfaces, it makes the kind of detail-oriented, integrative, and cross-cutting questions that motivated AIRI's research program possible to ask — and to answer — in one place.

The AI Risk Navigator was built by me, Spencer Michaels, during a Spring 2026 research fellowship at the Cambridge Boston Alignment Initiative.

Creating this website would not have been possible without the support and guidance of my mentors, Alexander Saeri and Peter Slattery. A huge thank you also goes to my research manager, Emre Yavuz, and my collaborators at AIRI, Simon Mylius, Ben Olsen, and Victoria Snorovikhina.

Finally, thank you to all of my CBAI friends (especially the Dana St residents) for making breaks, weekends, and the entire CBAI fellowship such an incredible time.

The datasets that power this tool represent years of research by the AIRI team and its collaborators. Any errors in their presentation here are my own.

— Spencer