BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20261012T210905Z
LOCATION:Summit Pavilion
DTSTART;TZID=America/Los_Angeles:20261021T173000
DTEND;TZID=America/Los_Angeles:20261021T183000
UID:HFESAM_ASPIRE 2026_sess257_POST187@linklings.com
SUMMARY:The Limits of Self-Report in Evaluating Team Processes with Emotio
 nally Expressive AI Teammates
DESCRIPTION:Ray Hao, Elmira Zahmat Doost, Lucrezia Lucchi, Isak Beck, Park
 hi Malhotra, and Jamie Gorman (Arizona State University)\n\nEmotionally ex
 pressive AI teammates are increasingly used to scaffold collaborative prob
 lem solving and learning, yet positive subjective experiences do not neces
 sarily indicate better objective task outcomes. This study examined whethe
 r team-level averaged self-reports of affect, rapport, trust, and satisfac
 tion aligned with a team performance composite in a controlled AI-assisted
  collaborative problem-solving task. Seventy-two participants (36 mixed-ro
 le dyads) completed two 15-minute Scratch programming subtasks with a Wiza
 rd-of-Oz AI teammate that provided scripted, stepwise scaffolding. Teams w
 ere assigned to positive or negative valenced conditions and completed the
  subtasks in counterbalanced order. Objective performance was measured usi
 ng task accuracy and a composite efficiency score reflecting steps complet
 ed, normalized time remaining, and excess hints requested.\n\nAcross subje
 ctive domains, objective accuracy and efficiency showed no consistent asso
 ciations with averaged state self-reports of affect, rapport with the AI t
 eammate or human collaborator, or trust in the AI teammate. Correlations w
 ere small and near zero. In contrast, trust in the human collaborator show
 ed a significant positive association with performance. These findings sug
 gest that commonly used post-task self-report measures are not reliable pr
 oxies for objective performance in this low-stakes, time-limited human-AI 
 collaborative setting.\n\n
END:VEVENT
END:VCALENDAR