Curtis J. Bonk

This exploratory analysis investigates the integration of ChatGPT in self-directed learning (SDL). Examining YouTube content creators’ language-learning experiences, the study builds on Song and Hill’s online SDL model. Thematic analysis of interviews with 19 YouTubers and related content reveals distinct constructs of ChatGPT‑integrated SDL, suggesting refinements to the SDL framework for the era of generative AI. The framework highlights interplay between learner traits and processes (local) and the evolving socio‑political‑cultural AI landscape (global). Findings point to ChatGPT’s potential for promoting self‑directed language learning and offer implications for learning technologies and AI‑facilitated SDL research.

Using a sequential mixed-methods design (n=384 survey; 10 interviews), this study investigates how postsecondary students use ChatGPT to support self-directed learning (SDL) for writing. Learners primarily used ChatGPT for brainstorming and idea generation. Initial motivations (e.g., curiosity, course requirements) often transformed into task motivation as benefits became evident. Participants generally managed their learning responsibly and employed various self‑management strategies; while survey responses indicated comparatively low self‑monitoring, interviewees described critical reflection and validation of ChatGPT output. Perceived improvement in writing skills was mixed. Implications for designing AI‑integrated learning experiences are discussed.

This mixed‑methods study (online survey of 84 participants and 10 interviews) examined self‑management skills needed for self‑directed language learning with mobile‑assisted technology (Duolingo). Specific, well‑defined learning goals—whether self‑initiated or structured by the platform—were associated with better performance. Daily-life distractions challenged time management. Effective SDLL required seeking human and material resources beyond a single tool, with resource authenticity critical to meaningful learning. The study offers implications for optimizing SDLL experiences in informal settings.

With generative AI tools becoming widespread, this mixed‑methods study (276 survey respondents; 11 interviews) explored how language learners in the U.S. self‑direct their learning with ChatGPT. Learners were motivated by flexibility and personalization that aligned materials with knowledge levels, interests, and goals. Inherent self‑monitoring skills helped learners use ChatGPT more effectively. The study suggests ways to design AI‑integrated learning that supports diverse learner needs and expands opportunities for self‑directed language learning.

Synthesizing a decade of studies on self‑directed online learners (SDOLs), this paper examines their goals, motivations, challenges, and resource preferences in open and MOOC contexts. Methods included surveys, interviews, focus groups, and content analyses, including work with a database of 3,000+ MOOC instructors. SDOLs seek autonomy, freedom, and choice; instructors are intrinsically motivated to contribute to open learning. The article provides instructional design guidelines to address SDOL needs and cultivate the next generation of SDOLs.

The R2D2 model—Read, Reflect, Display, and Do—organizes online learning activities to address diverse learner preferences and strengths. The model’s quadrants emphasize knowledge acquisition via readings and lectures, reflective writing and e‑portfolios, visual representation through timelines and concept maps, and hands‑on application via simulations and cases. R2D2 provides an accessible framework to design engaging, responsive online learning for varied learners.

Using an electronic conferencing system (COW) with preservice teachers discussing field‑experience cases, this study examined whether case‑based instruction and web conferencing foster high‑quality discourse and critical thinking outside class time. While the system supported participation, analysis showed discourse was often experiential and lacked consistent, well‑supported reasoning. The authors discuss implications for interactivity, motivation, task structure, and participant readiness in online learning.

This study analyzed online discussion in a graduate educational psychology course using the starter‑wrapper technique. Quantitative measures compared participation rates; qualitative content analysis (Henri, 1992) examined interaction patterns, social cues, cognitive and metacognitive elements, and depth of processing. Despite minimal required posting, messages were lengthy, cognitively deep, and indicative of a student‑oriented environment. Interaction became more complex over time, influenced by the weekly discussion starter. The paper proposes refinements to content‑analysis models for CMC.