Dan Parshall, Ph.D.

Bio

I'm a former physicist and data scientist, now working full-time on AI policy and governance. My Ph.D. was in condensed-matter physics, and I spent several years as a physicist (including at NIST) before leaving for the greener pastures of Data Science, working in aviation safety (MITRE), edtech (Hobsons/PowerSchool), entertainment (Whip Media), and IT (Lakeside Software). In 2026 I founded the Canary Institute as the public face for my policy work.

I'm an extrovert with a knack for explaining complex topics in simple terms, whether to executives, students, or members of Congress. Most of all, I get things done. Based in Washington, DC.

Currently

At the Canary Institute I've been writing on AI corrigibility and the economics of AI deployment. A few recent pieces:

• "Navigator Identity as a Mechanism for AI Corrigibility": working paper on Bilateral Constitutional AI (BCAI). BCAI treats alignment as a skill to be practiced, not a state to be installed (pdf, blog summary).
• "Measuring AI's Economic Reach": co-authored. Introduces the Cognitive–Deployment–Regulatory (CDR) framework, and finds that 50.5% of U.S. economy-weighted labor time already meets all three accessibility criteria for AI (blog, interactive).
• "The Biggest Bet in History": what the $2.5 trillion in hyperscaler infrastructure commitments implies about AGI timelines (link).
• RadVac policy brief (COVID-era): I compressed a 90-page open-source intranasal vaccine whitepaper into a 5-page brief, and pitched several Latin American health ministries (pdf).

Full research + physics publication list on Research.

Contact

dan@canaryinstitute.ai
GitHub · LinkedIn · CV (pdf, 47 kB)

h-index: 12 · Erdős number: 6.