BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:Zoë Hitzig | How People Use ChatGPT
X-WR-TIMEZONE:Pacific Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260520T131709Z
UID:tag:localist.com\,2008:EventInstance_52179619800629
DTSTART:20260309T190000Z
DTEND:20260309T200000Z
DESCRIPTION:Despite the rapid adoption of LLM chatbots\, little is known ab
 out how they are used. We approach this question theoretically and empiric
 ally\, modeling a user who chooses whether to complete a task herself\, as
 k the chatbot for information that reduces decision noise\, or delegate ex
 ecution to the chatbot.\n\nThe model—a rational inattention problem with
  a noisy delegation option—predicts that querying is favored for high-st
 akes\, high-context decisions\, while delegation is favored for routine\, 
 low-context tasks where the chatbot has a productivity advantage. Empirica
 lly\, we study the growth of ChatGPT’s consumer product from its launch 
 in November 2022 through July 2025\, combining usage logs with a privacy-p
 reserving pipeline that classifies a representative sample of conversation
 s. We document rapid global diffusion—reaching around 10% of the world
 ’s adult population—with work use concentrated among highly educated\,
  highly paid professionals and an increasing share of non-work use\, which
  has risen from 53% to more than 70% of all messages.\n\nConversations are
  dominated by “Practical Guidance\,” “Seeking Information\,” and 
 “Writing\,” and users with more complex\, knowledge-intensive jobs—m
 anagers\, highly educated\, and higher-paid professionals—are more likel
 y to use ChatGPT as a decision aide than as an agent\, consistent with the
  model’s predictions.
GEO:37.429987;-122.17333
LOCATION:Gates Computer Science Building\, 119
SUMMARY:Zoë Hitzig | How People Use ChatGPT
URL;VALUE=URI:https://events.stanford.edu/event/zoe-hitzig-how-people-use-c
 hatgpt
CATEGORIES:Class/Seminar
END:VEVENT
END:VCALENDAR