Tegan McCaslin

Berkeley, CA

Bio

Updated 03/23/26

Tegan McCaslin is a generalist researcher working at the intersection of AI forecasting, AI strategy, and AI governance. She began her research career at AI Impacts in 2018, where she studied neuroscience topics relevant to AI timelines, including brain architecture and neuron counts across species. She received multiple Long-Term Future Fund grants (2019, and subsequently) to pursue independent research into AI forecasting and strategy questions, exploring topics such as whether AI capability development parallels biological evolution and the tractability of long-term forecasting. She went on to join the Forecasting Research Institute (FRI) as a core founding team member alongside Phil Tetlock, focusing on improving the quality and decision-relevance of forecasting questions and the challenges of forecasting on long timescales. She co-authored FRI's report on Conditional Trees as a method for generating informative AI risk forecasting questions, and served as a mentor for the Epoch and FRI mentorship program for women and non-binary people interested in AI forecasting. More recently, she has expanded into AI governance work, contributing to the STREAM (ChemBio) framework at the Centre for the Governance of AI (GovAI), a standard for transparently reporting AI model evaluations of chemical and biological capabilities.