Speaking to the Machine: AI and Unspoken Lessons in Education

• In Event: Teaching and Learning in Sociology Section Open Call Paper Session

Abstract

Speaking to the Machine examines the pedagogical potential and social risks of artificial intelligence (AI) in higher education. As AI becomes increasingly embedded in educational contexts, this study asks two interrelated questions: How can AI enhance classroom instruction, and does its use reproduce or amplify existing social biases? Drawing on literature in active learning, Human-Centered AI, and algorithmic bias, the study situates AI as both an innovative instructional aide and a potential site of inequality.

The research centers on G𝜋T (Generative Practice Interview Trainer), a customizable AI-powered mock interview chatbot designed to integrate course content with professional skill development. Implemented across nine colleges at a diverse, minority-serving university during the 2025–2026 academic year, the project engaged ninety students over two semesters. Faculty tailored the chatbot to align with disciplinary learning goals, while students used it to simulate interviews based on real job postings. Data include interview transcripts, chatbot-generated scores, and faculty and student surveys.

Preliminary findings indicate strong pedagogical promise. G𝜋T enhances conceptual understanding, connects abstract course material to professional contexts, and increases student confidence and motivation. These benefits appear particularly meaningful for historically underrepresented and first-generation students, suggesting potential for narrowing equity gaps. However, transcript analysis may reveal subtle patterns of bias. The chatbot occasionally frames feedback differently based on perceived demographic cues, reinforcing gendered or racialized stereotypes in professional preparation. While AI can improve instructional practice and student outcomes, it must be critically evaluated to prevent the replication of structural inequalities.

As AI becomes integral to the workforce, unequal access to AI resources threatens to widen existing gaps. This study highlights the need for equitable AI education. It provides students with accessibility to build their AI skills, while promoting equity between course and career readiness within the higher education landscape.

Authors

• Edward Watson, California State University-Fullerton

• Alison Marzocchi

• Sunny Le

• Francisco Zepeda

• Matheus Guerrero