AI Assisted Surveys II:AI for the Survey Lifecycle
Credits:2
Grad Meth:
Reg, Aud
Prerequisite: SURV605: AI Assisted Surveys I.
Recommended: The students should have at least one paid subscription to advanced LLM models such as the latest Gemini model, ChatGPT, or Claude, for example. We will also experiment with open-weights models such as meta.ai and others.
Additional information: Weekly assignments and exercises will help the students practice what was learned during the lectures and sometimes the assignments will be to reproduce an already published paper with the latest LLM models and tools to compare its results to the original ones., , Finally, we will discuss the potential threats of AI agents filling out online surveys and the controversial topic of synthetic respondents.
Introduces the students to the latest literature and practices on AI assisted surveys., For each step of the survey process lifecycle, we will test how and how well, or not, AI tools can assist and sometimes replace the researcher in designing, fielding, analyzing, summarizing and documenting/archiving survey findings and datasets. , , The mental model we will use is to think about each survey task as potentially being assisted, or replaced by an AI agent/tool with the researcher in the loop. We will review the latest literature with practical examples of usage of AI tools at each step of the survey process.