Convergence Analysis

Convergence Analysis is an international AI existential risk strategy think tank with the mission of designing a safe and flourishing future for humanity in a world with transformative AI. The organization was originally founded as a research collaboration in existential risk strategy between David Kristoffersson and Justin Shovelain from 2017 to 2021, engaging a diverse group of collaborators. It was formally incorporated as a 501(c)(3) nonprofit in 2018. Through 2021 to 2023, the founders laid the foundation for a research institution and built a larger team, and in 2024 Convergence relaunched with a cross-functional team of academics and professionals.

Convergence operates three core research programs. The AI Clarity program builds a foundational series of sociotechnical reports modeling plausible and highly consequential AI scenarios, including the Threshold 2030 conference (co-organized with Metaculus and the Future of Life Institute) that brought together 30 leading economists, AI policy experts, and professional forecasters to evaluate AI's potential economic impacts. The Governance Recommendations program evaluates the most critical and neglected governance strategies, providing detailed analyses of specific AI governance proposals. The AI Awareness program disseminates research through books, podcasts, and public education.

In 2024, Convergence published over 470 pages across 19 articles and reports, provided consultation to the US Bureau of Industry and Security that directly informed proposed rules on dual-use AI reporting requirements, and had specific recommendations incorporated into the EU AI Act GPAI Code of Practice. Key team members include CEO and co-founder David Kristoffersson, co-founder and Chief Strategist Justin Shovelain, Dr. Justin Bullock (editor of the Oxford Handbook of AI Governance), and Dr. Christopher DiCarlo (author of Building a God and host of the All Thinks Considered podcast). Notable 2025 publications include The Manhattan Trap, A Global AGI Agency framework, the 200-page Threshold 2030 Full Report, and Pathways to Short AI Timelines.