ORCID

https://orcid.org/0009-0004-5225-8481

Keywords

AI-mediated feedback, feedback literacy, source credibility, MAIN model, communication education, speech outline, human-machine communication, formative feedback, instructor-designed AI

Subject Categories

Educational Psychology | Higher Education

Abstract

This study examined whether instructor-designed AI-mediated feedback and direct instructor feedback produced different outcomes when undergraduate students revised informative speech outlines. Using a within-subjects experimental design (n = 46), participants received both feedback types across two outline revision tasks under blinded conditions. AI-mediated feedback was delivered through Logos, a custom GPT system configured by me using the same rubric criteria, tonal guidelines, and relational message-design principles used for direct instructor feedback. Feedback delivery source served as the within-subjects independent variable. Outcomes included perceived helpfulness, perceived source credibility (competence, trustworthiness, and goodwill), self-reported feedback use, and changes in speech outline quality. Pre-existing feedback literacy disposition was examined as a moderating variable. No statistically significant differences were found between the two feedback delivery sources for any primary outcome. However, both conditions produced statistically significant improvements in outline quality. Feedback literacy did not produce a reliable relative advantage for either feedback source in this sample, although the restricted range of feedback-literacy scores may have limited the detection of differences. Supplementary exploratory analyses revealed that students who correctly identified the AI-mediated feedback as AI-generated rated it as significantly less credible and less helpful, supporting the MAIN model agency cue hypothesis, although their self-reported feedback use did not differ from that of students who did not correctly identify the source. Feedback literacy asymmetry also emerged, with stronger associations between literacy and perceptual outcomes in the direct instructor feedback condition than in the AI-mediated condition. Findings are discussed in relation to the MAIN model, the Computers Are Social Actors paradigm, source credibility theory, and feedback literacy theory, alongside ethical considerations related to disclosure, quality oversight, and communication education practice.

Completion Date

2026

Semester

Summer

Committee Chair

Nan Yu

Degree

Doctor of Philosophy (Ph.D.)

College

College of Sciences

Department

Nicholson School of Communication and Media

Format

PDF

Document Type

Dissertation

Language

English

Release Date

8-15-2028

Download

Available for download on Tuesday, August 15, 2028

Share

COinS
 

Accessibility Statement

This item was created or digitized prior to April 24, 2027, or is a reproduction of legacy media created before that date. It is preserved in its original, unmodified state specifically for research, reference, or historical recordkeeping. In accordance with the ADA Title II Final Rule, the University Libraries provides accessible versions of archival materials upon request. To request an accommodation for this item, please submit an accessibility request form.