Mark Warschauer
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Mark Warschauer has a total of 174 co-authors. Here are the top co-authors by number of shared publications:
- Tamara Powell Tate
University of California, Irvine, USA
23 shared publications - Sharin Jacob
Digital Promise Global, Washington, DC, USA
14 shared publications - Daniel Ritchie
School of Education, University of California, Irvine, Irvine, California, USA
10 shared publications - Santiago Ojeda-Ramirez
University of California
10 shared publications - Ying Xu
Graduate School of Education, Harvard University, Cambridge, Massachusetts, USA
7 shared publications - Dana Saito-Stehberger
University of California, Irvine, Irvine, CA, USA
6 shared publications
Latest Publications
- Open Access
- 15 October 2026
- Published by Center for Open Science in PsyArXiv
A Critical Epistemological Audit of Early Evidence Syntheses in Generative AI Educational Research: A Systematic Review and Meta-Analysis
The rapid emergence of ChatGPT in education was accompanied by an equally rapid proliferation of meta-analyses synthesizing its effects on educational outcomes. We conducted a systematic epistemological audit of the first wave of these meta-analyses, evaluating both their methodological and reporting quality and the primary evidence they synthesized. We included meta-analyses published after the release of ChatGPT that quantitatively synthesized its effects on educational outcomes. Five meta-analyses comprising 87 unique primary studies met our inclusion criteria. We also reconstructed their quantitative evidence base and conducted exploratory three-level random-effects meta-analyses to examine effect magnitude, heterogeneity, and robustness, and how these varied with theoretically and methodologically relevant characteristics. The five meta-analyses generally searched multiple databases, specified eligibility criteria, and examined heterogeneity and potential publication bias, but showed limitations in independent screening and coding, preregistration, open data and code, and transparent reporting. The underlying primary evidence likewise showed methodological and reporting limitations and substantial conceptual heterogeneity. Among 69 classifiable primary studies, 79.7% assessed performance exclusively, whereas only 11.6% assessed learning exclusively. The reconstructed meta-analysis yielded a positive average effect (g = 0.83, 95% CI [0.68, 0.97]), but heterogeneity was substantial (I² = 94.06%; 95% prediction interval [−0.75, 2.41]), and estimates were markedly attenuated in sensitivity analyses addressing small-study and selection effects. Although positive effects remained among stronger studies assessing learning, increasingly stringent quality criteria rapidly reduced the evidence base, limiting the certainty and generalizability of conclusions about learning effects. Overall, our findings caution against interpreting a single pooled estimate as a stable general effect of ChatGPT on learning and highlight the need for more rigorous, conceptually precise, transparent, and reproducible research.
- 15 October 2026
- Published by Center for Open Science in PsyArXiv
Open AccessA Critical Epistemological Audit of Early Evidence Syntheses in Generative AI Educational Research: A Systematic Review and Meta-Analysis
The rapid emergence of ChatGPT in education was accompanied by an equally rapid proliferation of meta-analyses synthesizing its effects on educational outcomes. We conducted a systematic epistemological audit of the first wave of these meta-analyses, evaluating both their methodological and reporting quality and the primary evidence they synthesized. We included meta-analyses published after the release of ChatGPT that quantitatively synthesized its effects on educational outcomes. Five meta-analyses comprising 87 unique primary studies met our inclusion criteria. We also reconstructed their quantitative evidence base and conducted exploratory three-level random-effects meta-analyses to examine effect magnitude, heterogeneity, and robustness, and how these varied with theoretically and methodologically relevant characteristics. The five meta-analyses generally searched multiple databases, specified eligibility criteria, and examined heterogeneity and potential publication bias, but showed limitations in independent screening and coding, preregistration, open data and code, and transparent reporting. The underlying primary evidence likewise showed methodological and reporting limitations and substantial conceptual heterogeneity. Among 69 classifiable primary studies, 79.7% assessed performance exclusively, whereas only 11.6% assessed learning exclusively. The reconstructed meta-analysis yielded a positive average effect (g = 0.83, 95% CI [0.68, 0.97]), but heterogeneity was substantial (I² = 94.06%; 95% prediction interval [−0.75, 2.41]), and estimates were markedly attenuated in sensitivity analyses addressing small-study and selection effects. Although positive effects remained among stronger studies assessing learning, increasingly stringent quality criteria rapidly reduced the evidence base, limiting the certainty and generalizability of conclusions about learning effects. Overall, our findings caution against interpreting a single pooled estimate as a stable general effect of ChatGPT on learning and highlight the need for more rigorous, conceptually precise, transparent, and reproducible research.
- 10 September 2026
- Published by Springer Nature
- p. 1-18
The Digital Divide: Manifestations, Proposed Solutions, and Future Implications in the Age of AI
- 10 September 2026
- Published by Springer Nature
- p. 1-18
The Digital Divide: Manifestations, Proposed Solutions, and Future Implications in the Age of AI
- Open Access
- 18 August 2026
- Published by Center for Open Science in EdArXiv
Generative AI as a Mediational Agent: Rethinking Learning in Sociocultural Theory
Generative AI challenges a foundational distinction in sociocultural theories of learning: the separation between mediational means and social interaction. Traditionally, tools such as language, writing, and educational technologies have been understood as mediating human activity, while learning and development arise through social participation with teachers, peers, and communities. Generative AI complicates this framework because it both mediates activity and generates context-sensitive, contingent contributions that shape ongoing interaction. This essay argues that existing descriptions of AI as either a tool or a collaborator are insufficient. Treating AI solely as a tool underestimates its interactional influence, while treating it as a collaborator risks attributing intentionality, accountability, and community membership that AI systems do not possess. To address this conceptual gap, the paper proposes the concept of the mediational agent: a responsive but non-accountable system that mediates human action while contributing explanations, critiques, questions, and suggestions to learning activity. Reconceptualizing generative AI in this way shifts attention from technological capability to forms of participation, highlighting the need for human-first habits that preserve learners’ judgment, agency, and responsibility in AI-mediated learning.
- 18 August 2026
- Published by Center for Open Science in EdArXiv
Open AccessGenerative AI as a Mediational Agent: Rethinking Learning in Sociocultural Theory
Generative AI challenges a foundational distinction in sociocultural theories of learning: the separation between mediational means and social interaction. Traditionally, tools such as language, writing, and educational technologies have been understood as mediating human activity, while learning and development arise through social participation with teachers, peers, and communities. Generative AI complicates this framework because it both mediates activity and generates context-sensitive, contingent contributions that shape ongoing interaction. This essay argues that existing descriptions of AI as either a tool or a collaborator are insufficient. Treating AI solely as a tool underestimates its interactional influence, while treating it as a collaborator risks attributing intentionality, accountability, and community membership that AI systems do not possess. To address this conceptual gap, the paper proposes the concept of the mediational agent: a responsive but non-accountable system that mediates human action while contributing explanations, critiques, questions, and suggestions to learning activity. Reconceptualizing generative AI in this way shifts attention from technological capability to forms of participation, highlighting the need for human-first habits that preserve learners’ judgment, agency, and responsibility in AI-mediated learning.
- Open Access
- 6 August 2026
- Published by Elsevier in SSRN Electronic Journal
Characterizing AI Support Patterns in Child-AI Dialogue During a Science-Themed Planet-Design Activity: A Computational Interaction Analysis
This study characterizes how a generative AI voice assistant’s support functions co-occurred with young children’s questions during a science-themed planet-design activity. The sample comprised 216 children aged 5–7 participating in 108 dyadic sessions. Children’s questions and AI responses were manually coded, with AI responses classified as cognitive, metacognitive, or motivational support. Information-seeking questions predominated, and cognitive support was the most frequent AI response category. Across sessions, information-seeking questions were followed by cognitive support in 529 of 658 transitions (.80). Density-based sequence clustering identified five provisional session profiles characterizing variation around this shared interactional backbone, while 36.1% of sessions remained unclustered. Children generally reported high enjoyment, low perceived difficulty, and strong preferences for future AI use. Exploratory regressions with standard errors clustered by dyad yielded several nominal associations with children’s reported experience; after false-discovery-rate adjustment, only daily-life questions and partner age remained associated with enjoyment. The findings provide a descriptive account of recurring question–support sequences and demonstrate how computational interaction analysis can characterize child-AI dialogue in early science-themed activities.
- 6 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessCharacterizing AI Support Patterns in Child-AI Dialogue During a Science-Themed Planet-Design Activity: A Computational Interaction Analysis
This study characterizes how a generative AI voice assistant’s support functions co-occurred with young children’s questions during a science-themed planet-design activity. The sample comprised 216 children aged 5–7 participating in 108 dyadic sessions. Children’s questions and AI responses were manually coded, with AI responses classified as cognitive, metacognitive, or motivational support. Information-seeking questions predominated, and cognitive support was the most frequent AI response category. Across sessions, information-seeking questions were followed by cognitive support in 529 of 658 transitions (.80). Density-based sequence clustering identified five provisional session profiles characterizing variation around this shared interactional backbone, while 36.1% of sessions remained unclustered. Children generally reported high enjoyment, low perceived difficulty, and strong preferences for future AI use. Exploratory regressions with standard errors clustered by dyad yielded several nominal associations with children’s reported experience; after false-discovery-rate adjustment, only daily-life questions and partner age remained associated with enjoyment. The findings provide a descriptive account of recurring question–support sequences and demonstrate how computational interaction analysis can characterize child-AI dialogue in early science-themed activities.
- Open Access
- 21 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 989-995
Designing Conversational Agents for Young Children: Comparing Disembodied, Embodied, and Customizable Agents
- 21 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 989-995
Open AccessDesigning Conversational Agents for Young Children: Comparing Disembodied, Embodied, and Customizable Agents
- Open Access
- 26 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 436-446
From Wandering to Collaboration: Discourse Patterns in Middle School Generative AI Use
- 26 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 436-446
Open AccessFrom Wandering to Collaboration: Discourse Patterns in Middle School Generative AI Use
- Open Access
- 22 April 2026
- Published by Center for Open Science in EdArXiv
Editor, Tutor, or Thinking Partner: How Students Position AI in the Writing Classroom
Collaboration in writing development and revision is a longstanding tradition shown to have positive impacts on both writing process and product across fields and disciplines. However, this tradition is undergoing a marked shift as generative AI tools transform what it means to collaborate, receive feedback, and iterate with a partner. While these processes are well established in human-to-human contexts, far less research examines collaboration between humans and AI. This study engages traditional collaborative writing research with human-computer interaction scholarship to explore how students mimic established role designations and interaction patterns in their engagement with generative AI systems during an AI-integrated upper-division writing course. Findings indicate that students both explicitly and implicitly assigned their AI partners familiar collaborative roles such as tutor/teacher, peer editor, and thinking partner, which shaped not only their engagement with the tool but also their perceptions of the writing process.
- 22 April 2026
- Published by Center for Open Science in EdArXiv
Open AccessEditor, Tutor, or Thinking Partner: How Students Position AI in the Writing Classroom
Collaboration in writing development and revision is a longstanding tradition shown to have positive impacts on both writing process and product across fields and disciplines. However, this tradition is undergoing a marked shift as generative AI tools transform what it means to collaborate, receive feedback, and iterate with a partner. While these processes are well established in human-to-human contexts, far less research examines collaboration between humans and AI. This study engages traditional collaborative writing research with human-computer interaction scholarship to explore how students mimic established role designations and interaction patterns in their engagement with generative AI systems during an AI-integrated upper-division writing course. Findings indicate that students both explicitly and implicitly assigned their AI partners familiar collaborative roles such as tutor/teacher, peer editor, and thinking partner, which shaped not only their engagement with the tool but also their perceptions of the writing process.
- 24 February 2026
- Published by Taylor & Francis
- p. 509-523
CALL and AI Literacy
This chapter examines the intersection of technology and English for Academic Purposes (EAP), with a focus on the transformative role of artificial intelligence (AI). It begins with a review of the historical development of Computer-Assisted Language Learning (CALL), highlighting AI as its most recent advancement and exploring its implications for EAP. The chapter then addresses key issues in the integration of AI into EAP, discussing both its affordances and challenges. To illuminate current research trends, we summarize key findings on how AI technologies—such as intelligent tutors, adaptive systems, and virtual reality (VR)—contribute to three major areas of EAP: academic reading, vocabulary, and writing. We also introduce an AI-based writing platform, PapyrusAI, and highlight its specific affordances for supporting multilingual writers in academic settings. The chapter concludes with practical recommendations and discusses future directions for technology-supported EAP.
- 24 February 2026
- Published by Taylor & Francis
- p. 509-523
CALL and AI Literacy
This chapter examines the intersection of technology and English for Academic Purposes (EAP), with a focus on the transformative role of artificial intelligence (AI). It begins with a review of the historical development of Computer-Assisted Language Learning (CALL), highlighting AI as its most recent advancement and exploring its implications for EAP. The chapter then addresses key issues in the integration of AI into EAP, discussing both its affordances and challenges. To illuminate current research trends, we summarize key findings on how AI technologies—such as intelligent tutors, adaptive systems, and virtual reality (VR)—contribute to three major areas of EAP: academic reading, vocabulary, and writing. We also introduce an AI-based writing platform, PapyrusAI, and highlight its specific affordances for supporting multilingual writers in academic settings. The chapter concludes with practical recommendations and discusses future directions for technology-supported EAP.
- Open Access
- 17 February 2026
- Published by Association for Computing Machinery (ACM)
- p. 379-385
Boosting Coding Confidence in Elementary Students: The Impact of ELA-Integrated Computational Thinking Curriculum
- 17 February 2026
- Published by Association for Computing Machinery (ACM)
- p. 379-385
Open AccessBoosting Coding Confidence in Elementary Students: The Impact of ELA-Integrated Computational Thinking Curriculum
- 1 December 2025
- Published by EUROKD Egitm Danismanlik Group in Language Teaching Research Quarterly
- Vol. 51, 409-431
Bringing New Literacies into the Secondary Classroom: A Case Study of Synchronous Collaborative Writing with Google Docs
Despite the increasingly widespread adoption of collaborative technology in K-12 writing instruction, little is known about how these new literacy practices shape students’ academic writing experiences with synchronous online collaboration, particularly within the context of long-term curricular integration. Addressing this gap, this study examines middle school students' in-class use of Google Docs for synchronous collaborative writing over the course of a year, exploring how technology reshapes learning experiences and opens new pathways for interaction and literacy practices as suggested by Chapelle (2007). Based on qualitative analyses of multiple data sources—including student and teacher interviews, student reflections, and observations , we examined the perceived affordances and challenges of these new literacy practices, as well as the contextual factors that influence students’ collaborative experiences. Findings revealed several key benefits of synchronous academic writing, including enhanced writing fluidity, reflective verbal composing, and audience awareness, while also revealing challenges such as an overemphasis on efficiency, and tensions over textual ownership. We conclude by emphasizing the pedagogical implications of effectively integrating synchronous collaborative writing into secondary classrooms and beyond.
- 1 December 2025
- Published by EUROKD Egitm Danismanlik Group in Language Teaching Research Quarterly
- Vol. 51, 409-431
Bringing New Literacies into the Secondary Classroom: A Case Study of Synchronous Collaborative Writing with Google Docs
Despite the increasingly widespread adoption of collaborative technology in K-12 writing instruction, little is known about how these new literacy practices shape students’ academic writing experiences with synchronous online collaboration, particularly within the context of long-term curricular integration. Addressing this gap, this study examines middle school students' in-class use of Google Docs for synchronous collaborative writing over the course of a year, exploring how technology reshapes learning experiences and opens new pathways for interaction and literacy practices as suggested by Chapelle (2007). Based on qualitative analyses of multiple data sources—including student and teacher interviews, student reflections, and observations , we examined the perceived affordances and challenges of these new literacy practices, as well as the contextual factors that influence students’ collaborative experiences. Findings revealed several key benefits of synchronous academic writing, including enhanced writing fluidity, reflective verbal composing, and audience awareness, while also revealing challenges such as an overemphasis on efficiency, and tensions over textual ownership. We conclude by emphasizing the pedagogical implications of effectively integrating synchronous collaborative writing into secondary classrooms and beyond.