| Description |
*** IMPORTANT ***
For the most updated administrative course information (date changes, room changes etc) please always refer to the KSL page only and not to the Info page in ILIAS – the ILIAS infopage will not be updated!
***********************************
Prof. Dr. Jean-Michel Benkert
This course has limited slots (12) - please register in KSL
open from 01.-10.09.2026
Seminar dates:
Wednesday, 16.09. / 23.09. / 30.09.2026, 10.15-12.00hrs, A027, UniS (Lectures 1-3)
Wednesday, 07.10.2026, 10.15-11.00hrs, A027, UniS (written exam)
Wednesday, 14.10.2026, 10.15-12.00hrs, A027, UniS (Experimental Project launch)
Wednesday, 04.11.2026, 10.15-12.00hrs, A027, UniS (reflection + Many Analysts Project launch)
Wednesday, 02.12.2026, 10.15-12.00hrs, A027, UniS (Many Analysts group presentations)
Wednesday, 09.12.2026, 10.15-12.00hrs, A027, UniS (final session)
Artificial intelligence (AI) is rapidly transforming economies. At the same time, the tools we use to study economics — agentic AI assistants such as Claude Code and Codex — are themselves becoming powerful enough to change what it means to do (empirical) research in economics. This seminar takes both phenomena seriously.
The central question: how do we structure AI tools and workflows in economic research to capture the productivity gain without losing human capital? The course addresses this from three angles. First, three lectures develop an economic framework for delegation, verification, attention, and human capital; examine evidence on AI’s task-level effects; and ask how these effects aggregate through adoption, task composition, and labour-market adjustment. They also translate these lessons into effective AI-assisted workflows. Second, two projects contrast a regime with an externally checkable target (the Experimental Project, implementing an economic model in oTree) with a judgment-laden regime in which multiple defensible answers are possible (the Many Analysts Project, in which all students answer the same research question with the same dataset and compare approaches in randomly formed groups). Third, an explicit reflection session connects the project experience back to the course framework. |
User:
Öffentlicher Bereich
Support