BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.4.0.1//EN
TZID:Europe/Paris
X-WR-TIMEZONE:Europe/Paris
BEGIN:VEVENT
UID:181@isdm.umontpellier.fr
DTSTART;TZID=Europe/Paris:20260217T140000
DTEND;TZID=Europe/Paris:20260217T140000
DTSTAMP:20260825T125942Z
URL:https://isdm.umontpellier.fr/events/does-ai-help-humans-make-better-de
 cisions-a-statistical-evaluation-framework-for-experimental-and-observatio
 nal-studies/
SUMMARY:Does AI help humans make better decisions? A statistical evaluation
  framework for experimental and observational studies
DESCRIPTION:Online\n\nMachine Learning in Montpellier\, Theory &amp\; Pract
 ice - Melody Huang (Yale)\n\nThe use of Artificial Intelligence (AI)\, or 
 more generally data-driven algorithms\, has become ubiquitous in today's s
 ociety. Yet\, in many cases and especially when stakes are high\, humans s
 till make final decisions. The critical question\, therefore\, is whether 
 AI helps humans make better decisions compared to a human-alone or AI-alon
 e system. We introduce a new methodological framework to empirically answe
 r this question with a minimal set of assumptions. We measure a decision m
 aker's ability to make correct decisions using standard classification met
 rics based on the baseline potential outcome. We consider a single-blinded
  and unconfounded treatment assignment\, where the provision of AI-generat
 ed recommendations is assumed to be randomized across cases with humans ma
 king final decisions. Under this study design\, we show how to compare the
  performance of three alternative decision-making systems--human-alone\, h
 uman-with-AI\, and AI-alone. Importantly\, the AI-alone system includes an
 y individualized treatment assignment\, including those that are not used 
 in the original study. We also show when AI recommendations should be prov
 ided to a human-decision maker\, and when one should follow such recommend
 ations. We apply the proposed methodology to our own randomized controlled
  trial evaluating a pretrial risk assessment instrument. We find that the 
 risk assessment recommendations do not improve the classification accuracy
  of a judge's decision to impose cash bail. Furthermore\, we find that rep
 lacing a human judge with algorithms--the risk assessment score and a larg
 e language model in particular--leads to a worse classification performanc
 e.\n\nMachine Learning in Montpellier\, Theory &amp\; Practice
ATTACH;FMTTYPE=image/jpeg:https://isdm.umontpellier.fr/wp-content/uploads/
 2026/06/ml-mtp-gC78d5.png
CATEGORIES:ML MTP
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:STANDARD
DTSTART:20251026T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
END:VTIMEZONE
END:VCALENDAR