Week 6: Generative AI (e.g., ChatGPT) and Self-Directed Learning in Language Learning

Article 1: The Potential of AI in Writing Assessment: A Step Toward Objectivity or a Risk of Oversimplification?

Artificial intelligence (AI) is increasingly transforming education, and its role in assessing writing performance is a fascinating yet controversial topic. Jiang et al. (2023) explore the efficacy of large language models (LLMs) such as GPT-4, GPT-3.5, iFLYTEK, and Baidu AI Cloud in evaluating writing accuracy in second-language learners. While the study provides compelling evidence that AI can match human grading in precision, it also raises essential questions about the reliability, bias, and long-term implications of AI-powered assessment tools.

One of the key takeaways from this research is the comparison between human grading and AI grading. AI systems demonstrated robust accuracy in evaluating sentence-level and T-unit-level errors, suggesting that they could be instrumental in handling large-scale assessments efficiently. Given the well-documented inconsistencies in human grading due to subjectivity, fatigue, and bias, AI presents an attractive alternative. However, I wonder whether AI can truly grasp the nuances of contextual appropriateness, cultural implications, and creative expression in writing—elements that human graders are more attuned to.

Another fascinating aspect of this study is the potential for AI-assisted formative assessment. Since AI can provide immediate feedback, it could support self-regulated learning and help students improve their writing in real time. Yet, I remain skeptical about whether these models can distinguish between minor errors and deep structural issues in writing. Will students become overly reliant on AI-generated feedback rather than developing their own critical writing skills?

Additionally, the study highlights discrepancies between AI and human ratings. While AI might achieve high precision in identifying grammatical errors, human graders are still superior in holistic evaluation, such as assessing coherence, argument strength, and rhetorical effectiveness. Perhaps the future lies in a hybrid approach—leveraging AI for initial assessments while retaining human judgment for higher-order evaluation.

This article aligns well with my ongoing research interest in assessment literacy—specifically, how teachers can integrate AI in formative assessment without diminishing authentic learning experiences. Could AI tools be optimized to provide feedback on conceptual and structural elements, rather than just grammar and syntax? Would this shift AI’s role from a grader to a learning assistant?

As we embrace AI in education, we must ask: How can AI be leveraged to support—not replace—the essential human aspects of writing instruction? Should we trust AI to grade high-stakes writing assessments, or does it risk reducing language learning to a checklist of errors?

The promise of AI in writing assessment is undeniable, but the conversation must continue about ensuring that AI tools enhance, rather than dictate, educational practices.

 

Article 2: Empowering Language Learners Through AI: ChatGPT as a Tool for Educational Equity

The integration of artificial intelligence (AI) into education is transforming the way students learn, particularly in language acquisition. The study by Li, Li, and Cho (2023) highlights how ChatGPT can serve as an effective tool for improving Chinese writing skills among students from low-income backgrounds. This research is particularly intriguing because it explores the equity implications of AI in education, addressing a pressing concern—how technology can help bridge learning gaps for students with limited resources.

One of the key findings that stood out is how ChatGPT supports second-language learners (CLLs) by providing real-time feedback, correcting errors, and enhancing sentence structure. These features align well with the growing need for personalized and accessible learning tools. Traditional tutoring services are often financially out of reach for many low-income students, and AI presents a cost-effective alternative. However, this also raises ethical concerns: Should we rely on AI tools to replace human instruction, or should they serve as supplementary aids?

Another interesting aspect of this study is its single-case experimental design, which effectively captures individual progress. The use of an ABA design (baseline-intervention-reversal) strengthens the credibility of the findings by demonstrating how student performance fluctuates with and without AI assistance. It is fascinating to see that each participant showed improvement in their writing scores during the intervention phase. However, I wonder whether long-term reliance on ChatGPT could hinder independent language processing skills—could students become overly dependent on AI-generated corrections instead of developing their own grammatical awareness?

This study also ties into the broader debate on AI's role in formative assessment. If ChatGPT can provide immediate feedback, could it be leveraged not just for writing improvement but also for self-assessment and progress tracking? Teachers could potentially use AI-generated analytics to tailor instruction to individual student needs. However, this raises a critical question: How do we ensure that AI-generated feedback aligns with pedagogical best practices and does not misguide learners?

In conclusion, while ChatGPT offers great promise in making language learning more equitable, it should be used thoughtfully and ethically. AI can enhance formative assessment, increase engagement, and reduce educational inequality, but human oversight remains essential. Looking forward, it would be interesting to explore how AI can be optimized to foster deeper comprehension rather than just surface-level corrections.

What do you think? Can AI-powered tools like ChatGPT replace traditional tutoring, or should they remain supplementary aids? How can we ensure students use them effectively without becoming overly reliant? Let’s discuss it!

 

Article 3: The Power of Self-Directed Language Learning: Duolingo’s Role in Out-of-Class Learning

The study by Li and Bonk (2023) provides valuable insights into the world of self-directed language learning (SDLL) and how learners engage with technology-based tools like Duolingo outside formal educational settings. The findings highlight a shift in motivation and learning autonomy, reinforcing the growing role of mobile-assisted language learning (MALL) in self-paced education.

One key takeaway from the study is that learners actively manage their learning resources, relying not only on Duolingo but also on external materials and personal strategies to self-monitor progress. This aligns with metacognitive strategies, where learners engage in self-evaluation and judgment of their knowledge gaps, reinforcing the importance of learner autonomy in digital learning environments. This resonates with my own experience—while apps like Duolingo provide structure, real mastery often comes from diverse exposure, such as engaging with native speakers, watching foreign media, or practicing writing skills beyond the app’s scope.

Another fascinating finding is that intrinsic motivation plays a bigger role than extrinsic rewards. Unlike traditional language learning in academic settings, where grades and certifications drive engagement, Duolingo learners are driven by curiosity, cultural interest, travel aspirations, and cognitive stimulation (brain training). This raises a compelling question: Would formal education benefit from incorporating more intrinsic motivational elements to enhance engagement and long-term retention?

However, one area that warrants further exploration is Duolingo’s effectiveness in fostering communicative competence. While it is an excellent tool for vocabulary acquisition and basic grammar reinforcement, it lacks meaningful conversational practice. This gap in functional language use leads to an important consideration—can AI-driven language models enhance Duolingo by providing real-time, adaptive conversational feedback?

Final Thoughts & Discussion

Li and Bonk’s study underscores the power of learner autonomy and digital tools in language learning, particularly in informal settings. However, how can SDLL be optimized to ensure learners develop well-rounded language proficiency rather than just gamified engagement? Should language apps integrate AI-based conversational features or collaborate with live tutors to provide a more immersive learning experience?

References

Jiang, Z., Xu, Z., Pan, Z., He, J., & Xie, K. (2023). Exploring the role of artificial intelligence in facilitating assessment of writing performance in second language learning. Languages, 8(4), 247. https://doi.org/10.3390/languages8040247

Li, X., Li, B., & Cho, S.-J. (2023). Empowering Chinese language learners from low-income families to improve their Chinese writing with ChatGPT’s assistance afterschool. Languages, 8(4), 238. https://doi.org/10.3390/languages8040238

Li, Z., & Bonk, C. J. (2023). Self-directed language learning with Duolingo in an out-of-class context. Computer Assisted Language Learning. https://doi.org/10.1080/09588221.2023.2206874

 

Comments

Popular posts from this blog

Introduction: About myself