SELECTED THINGS I'M WORKING ON
Agentic inequality (with Omer Bilgin, Oxford, Iason Gabriel, DeepMind, and Lewis Hammond, Cooperative AI Foundation). How agentic AI systems risk concentrating economic and political power, and the governance responses available. Accepted at Nature Machine Intelligence.
Unlocking AI usage data for labour market research (with co-authors at the University of Oxford, the Stanford Digital Economy Lab, and the Centre for the Governance of AI). How aggregated usage data from frontier labs can track adoption at task and population scale, and how access governance should be designed across developers, governments, and statistical agencies.
Quantitative empirical work on AI adoption and digital payments. Applied statistical methods to estimate economic effects - from country-level studies of AI adoption and productivity through to impact evaluations of digital payments using granular transactions data. (e.g. a G20/BIS working paper)
Frontier AI governance in emerging economies. How developing countries should approach the governance of frontier AI, given different state capacities and exposures to its economic effects. Recently presented at the ITU's TDAG meeting in Geneva.
Building automation workflows and delivering LLM training for tech policy consulting. I have been building end-to-end workflows: tailored policy monitoring, templated slide production, proposal development, etc. I also have designed and delivered many hours of training for all the consulting staff in my company.