The emergence of generative artificial intelligence (GenAI) has introduced both promising opportunities and legitimate concerns in the educational sphere. The use of artificial intelligence (AI) in education has gained significant attention in recent years, with a growing number of institutions and organizations exploring the potential benefits of AI-driven technologies (Dwivedi et al., 2021; Su & Yang, 2022). GenAI challenges educators to ensure that students benefit from AI in ways that enhance essential skills. Rather than limiting AI’s use in education, educators should seek proactive strategies to adapt and incorporate its capabilities thoughtfully (Chang et al., 2023).

This research project investigates the effective integration of generative AI to enhance students' metacognitive skills. As technology becomes more prevalent in schools, it is essential to integrate it in ways that empower learners to navigate the complexities of the digital age while preserving the critical thinking, collaboration, and self-regulation skills needed to become successful 21st-century learners.

Aligned with this purpose, the objectives of this research paper are to:

• Define metacognitive, self-regulated learning, sociocultural, and connectivism theories
• Examine how generative AI interacts with students’ metacognitive processes and self-regulation
• Identify effective strategies for utilizing generative AI in schools

Motivation of the Study

Before undertaking this research, my understanding of artificial intelligence (AI) was limited, and the prospect of incorporating it into my teaching felt intimidating. However, this challenge inspired me to explore how AI could enhance my students’ thinking skills and help me implement it effectively in the classroom. As a teacher of a Core Literature/Language Arts class, I work with students who need additional support in critical thinking, self-regulated learning, and executive function. This experience has been a driving force behind my interest in leveraging AI to address these specific needs.

Furthermore, the integration of AI into our district’s practices this year provided an additional layer of motivation. Beyond classroom instruction, AI has proven to be an invaluable tool for analyzing data more efficiently and accurately, as well as streamlining the preparation of school legal documents, such as Individualized Education Programs (IEPs). These applications have shown me the broad potential of AI to support both students and teachers in meaningful ways.

Ultimately, my goal is to create a more dynamic and engaging learning environment. Schools should not feel like a monotonous routine but rather an experience filled with variety and creativity. When used thoughtfully, AI has the potential to be an incredibly powerful tool, offering opportunities to design activities that support and challenge students while fostering their metacognitive skills.

Personal Motivation

I have always been drawn to technologies that enhance productivity, creativity, and engagement. As someone who grew up as a student in the 21st century, I’ve had access to various digital tools designed to enrich learning. However, I often reflect on how my own educational experiences could have been more meaningful and efficient if I had utilized the right tools at the right time.

The emergence of generative AI feels like a game-changer. It offers a unique opportunity to supplement critical thinking and creativity in ways that were previously unimaginable. Looking back, I can see how tools like generative AI could have helped me produce better work and explore topics more deeply. These reflections have fueled my passion for ensuring that today’s students have access to resources that empower them to learn more effectively and meaningfully.

 

Metacognition - A Brief Explainer [Video]. Education Endowment Foundation. (2023, March 4). YouTube. https://www.youtube.com/watch?v=sAik_RQY_Dg

What is Metacognition?

The research on metacognition has been going on for several decades. The concept of metacognition has been around before the term was coined by John Flavell in 1971. Flavell’s work laid the foundation for understanding how individuals can become aware of and control their cognitive processes to improve learning and problem-solving. He defined metacognition as "cognition about cognition" or "thinking about thinking” (Flavell, 1976, p. 232). The video suggests that explicit teaching of metacognition into educators' daily habits can support students to be more aware of their own learning and thinking, which can aid them to become effective and independent learners. Applying metacognition into students' learning methods can promote a positive outlook towards school and thus increase academic performance. Metacognition is essential in learning because students need to be aware of the demands of their tasks and goals, including how to use strategies and what is required to meet the demands of learning tasks.

Components of Metacognition

Figure 1. Components of Metacognition [Image]. Stanton, J. D., Sebesta, A. J., & Dunlosky, J. (2021). Fostering metacognition to support student learning and performance. CBE—Life Sciences Education, 20(2). https://doi.org/10.1187/cbe.20-12-0289

In his 1976 article and in Figure 1, Flavell recognized that metacognition consisted of both metacognitive knowledge and metacognitive regulation. According to Stanton, Sebesta, and Dunlosky (2021), metacognitive knowledge includes what you know about your own thinking and what you know about strategies for learning. The components under it are declarative knowledge, procedural knowledge, and conditional knowledge.

• Declarative knowledge, also known as person variable, involves knowing about yourself as a learner, the demands of the task, and what learning strategies exist.
• Procedural knowledge or task variable involves knowing how to use learning strategies.
• Conditional knowledge, also known as strategy variable, involves knowing when and why to use particular learning strategies.

On the other hand, metacognitive regulation involves the actions you take in order to learn. Its components include planning, monitoring, and evaluating.

• Planning includes selecting appropriate strategies and resources before starting a task.
• Monitoring is tracking one’s understanding and performance during the task.
• Evaluating is reflecting on assessing the effectiveness of strategies used, the success of the learning process, and whether goals have been achieved.

Other Theories Linked to Metacognition

There are several educational theories that are closely linked to metacognition including Zimmerman’s self-regulated learning theory, Vygotsky’s constructivist theory, and Siemens connectivism. Developed by educational psychologist, Barry Zimmerman, the self-regulated learning model explains how students can take control of their own learning processes through a cycle of planning, monitoring, and reflection. Zimmerman’s learning model includes three main phases namely forethought phase, performance phase, and self-reflection phase.

Zimmerman's Self-Regulated Learning

Figure 2. Phases and processes of self regulation [Image]. Zimmerman, B. J., & Moylan, A. R. (2009). Self-regulation: Where metacognition and motivation intersect. In D. J. Hacker, J. Dunlosky, & A. C. Graesser (Eds.), Handbook of metacognition in education (pp. 299–315).

As shown in Figure 2 above, the forethought phase refers to processes and beliefs that occur before efforts to learn. An example of this is a student setting a goal before starting a project, assessing their strengths to improve time management, and breaking down the project into steps. After the forethought phase comes the performance phase, which refers to processes that occur during behavioral implementation. For instance, a student tracks their time and or might use tools to help with focusing. The student also applies strategies and adjusts approaches as needed. Lastly, self-reflection refers to processes that occur after each learning effort. An example of self-regulated learning is when a student reflects whether their strategy and time management were effective and how they can improve their performance in the future. All the aforementioned skills can help enhance metacognition. Each phase contributes to the next, as self-reflection on one task often leads to an improvement in the future steps.

Vygotsky's Sociocultural Theory of Cognitive Development

Vygotsky's Sociocultural Theory [Video]. YouTube. (2023, November 26). YouTube. https://www.youtube.com/watch?v=gPKV2f7uDAo

As shown from the video, Vygotsky’s sociocultural theory of learning emphasizes the role of language and social interaction in cognitive development. Vygotsky’s theory comprises concepts such as The More Knowledgeable Other (MKO), the Zone of Proximal Development, scaffolding, and private speech. Vygotsky’s sociocultural theory supports metacognitive theory by highlighting the method of scaffolding, social and cultural aspects, and the use of internal dialogue. His theory also serves as a rich foundation for understanding metacognition, which augments learners’ acquisition of knowledge by developing their ability to think about their thinking and regulate their cognitive processes effectively.

The More Knowledgeable Other (MKO)

The more knowledgeable other (MKO) refers to someone who has a better understanding or higher skills level than the learner in a particular task or concept. Abtahi (2016) suggests that tools themselves can function as MKO that guide learners’ thinking and actions. The designs and affordances of tools can structure learning experiences, creating a zone of proximal development (ZPD) where learners, through their interactions with these tools, can achieve more than they could independently. This notion was supported by Puntambekar and Hübscher (2005), who discuss the use of curricula, software tools, and other resources as forms of scaffolding.

In the digital age, technology can facilitate MKO interactions, even in remote or online learning environments. Educational platforms and software can incorporate features that allow for collaborative learning and guidance (Cuofano, 2014).

Zone of Proximal Development

Figure 3. Model of Zone of Proximal Development [Image]. Practical Psychology. (2020, March). Zone of Proximal Development (Definition + Examples). Retrieved from https://practicalpie.com/zone-of-proximal-development/.

As shown in Figure 3, Vygotsky (1978) views the zone of proximal development as the area where the most sensitive instruction or guidance should occur, enabling the child to develop skills they will later use independently, thus fostering higher mental functions.

The concept of zone of proximal development (ZPD) was developed during the late 1920s and elaborated progressively until Vygotsky’s death in 1934. In Mind in Society: The Development of Higher Psychological Processes, Vygotsky defined the ZPD as “the distance between the actual development level as determined by independent problem solving and the level of potential development as determined through problem solving under adult guidance or in collaboration with more capable peer” (p. 86).

Roosevelt (2008) holds that the main goal of education from Vygotskian perspective is to keep learners in their own ZPDs as often as possible by giving them interesting and culturally meaningful learning and problem-solving tasks that are slightly more difficult than what they do alone, such that they will need to work together either with another, more competent peer or with a teacher or adult to finish the task.

Scaffolding

Vygotsky introduced scaffolding as an instructional strategy where a teacher or more knowledgeable peer provides support to help a learner achieve tasks within their Zone of Proximal Development (ZPD)—the range of tasks they can complete with assistance but cannot yet perform independently (Vygotsky, 1978).

The use of scaffolding supports the theory of metacognition by exposing students to effective learning strategies modeled by a teacher or a peer. This encourages students to reflect on and apply their strategies in different situations which can lead to a more conscious learning process. Scaffolding also promotes regulation of learning as students learn to take over the process initially modeled for them.

Vygotsky refers to this as cooperative or collaborative dialogue. The child seeks to understand the actions or instructions provided by the tutor (often the parent or teacher) and then internalizes the information, using it to guide or regulate their performance.

Private Speech

According to Vygotsky (1978), the key aspects of cultural tools are the language, symbols, physical tools, and social practices within a culture that shape thinking and learning. He believed that the concept of cognitive development is not only an individual process but also deeply influenced by the social and cultural context, mediated by tools and practices shared within that culture. Vygotsky considered language as the most important tool as it serves as a way to communicate, plan, reflect, and regulate behavior. Through social interaction, students internalize language and this forms as a basis of metacognition.

In Thought and Language (1934), Lev Vygotsky examined how language is central to cognitive development, particularly through its role in shaping thought processes and enabling self-regulation. Vygotsky argued that language is not only a means of communication but also a powerful tool for thinking, problem-solving, and learning. Private speech starts around age 2-7 to guide their actions. It serves as a way for children to plan, organize, and regulate their behavior, playing a crucial role in developing self-regulation, metacognition, and problem-solving skills.

Connectivism

Connectivism: A Networked Learning Theory for the Digital Age [Video]. YouTube. (2023a, September 21). YouTube. https://www.youtube.com/watch?v=ezw6_XHrwks

According to an article by Siemens (2005), connectivism is a modern learning theory that posits learning occurs through the process of creating connections among various networks of information, people, and resources. It was introduced by George Siemens and Stephen Downes in the early 2000s as an approach to understanding how learners acquire knowledge in the digital age, particularly in a world of rapidly expanding information and advanced technology. 

From the video, connectivism claims that the more nodes we connect, the more we acquire knowledge. Sources of information are not only limited to books and people but also digital tools such as websites, artificial intelliegence and social media platforms. Connectivism also theorizes the need for learners to learn from different sources and platforms as it aids them to have diverse knowledge about the world. This supports the notion that connectivism provides a framework that align with modern and networked learning environments. 

Connectivism and metacognition intersect in the way both encourage learners to be more self-aware and reflective about their learning processes. By combining both theories, learners can focus on networks and technology while engaging in metacognitive activities such as planning, monitoring, and evaluating learning. Applying such theories in this digital age can ensure that students are not just passive users of AI but active participants in their learning. 

Critics on Metacognition

In the research study Why Metacognition Is Not Always Helpful, Norman (2020) discusses why metacognition is not always helpful. It centers around three suggestions, namely, that (1) metacognition may actively interfere with task performance, (2) the costs of engaging in metacognitive strategies may outweigh the benefits, and that (3) metacognitive judgments or feelings involving a negative self-evaluation may detract from psychological well-being.

1. Metacognition May Actively Interfere With Task Performance

Concurrent verbalization of metacognition is when someone verbalizes their thoughts while simultaneously performing a task. The negative effect of verbalization on cognitive task performance is commonly referred to as “verbal overshadowing” (Yamada, 2009). Verbal overshadowing is a cognitive phenomenon where verbally describing an experience, such as a visual memory or a specific task, can interfere with one’s ability to accurately recall or perform that experience later (Schooler & Schooler, 1990). Verbalizing or describing an experience or skill can sometimes disrupt self-monitoring, a key aspect of metacognition where individuals assess their own understanding or performance. Verbal overshadowing shows that while language is a powerful tool for metacognitive reflection, it can also disrupt certain memories or skills, especially those rooted in nonverbal or procedural knowledge.

2. The Costs of Engaging in Metacognitive Strategies May Outweigh the Benefits

With the metacognitive strategies being learned either through explicit instruction or implicitly through everyday experiences, the intentional application of a metacognitive strategy could be demanding in terms of time and cognitive resources.

3. Metacognitive Judgments or Feelings Involving a Negative Self-Evaluation May Detract From Psychological Well-Being

Metacognitive beliefs (i.e., a form of metacognitive knowledge) may address a person’s evaluation of their own abilities and self-worth (Tarricone, 2011). For instance, a person may—correctly or incorrectly—assume that they are less able/talented than other people when it comes to some cognitive ability. This could in turn lower the person’s self-esteem and self-efficacy and thereby reduce their efforts in and motivation for trying to do their best on a certain cognitive task.

According to Veenman and Afflerbach (2006), one challenge that poses among educators is measuring metacognition. Assessing metacognitive skills can be complex, as they involve self-reporting and introspection, which are not always reliable. It is possible that self-reported metacognitive measures could inaccurately reflect actual metacognitive abilities or behavior. Measurement of metacognition is naturally difficult because metacognition is not an explicit behavior. With this, there is a continuous need for tools to measure metacognition. Schraw (2009) points out the difficulty of measuring metacognition and states that a single method that enables simultaneous connection to metacognition processes and allows for measurement of all of these processes alone does not exist.

Brown A.L (1987) recognized the advantages of metacognition, however, it also comes with drawbacks such as teaching challenges and unequal benefits for all learners. Brown argued that metacognitive skills are difficult to teach effectively, especially for younger children who may lack the cognitive maturity to engage in complex self-reflection or self-regulation. Metacognition requires higher-order thinking and such learning can be developmentally inappropriate for some learners. Another drawback Brown identified is the cognitive load that comes with metacognitive tasks. Engaging in metacognitive activities like planning, monitoring, and evaluating can be mentally taxing, especially when a task is already challenging, which could possibly lead to diminished performance.

Limitations of metacognition on explaining complex learning behaviors in digital environments include:

• Digital Distractions - multitasking and deliberate switching between tabs can make it challenging for learners to apply metacognitive skills
• Cognitive Overload - the abundance of information students get from digital tools can be overwhelming for students which can be challenging for students to engage in metacognitive activities
• Limited Social Interaction - metacognitive theory primarily focuses on individual thinking and self-regulation

To address these limitations, constructivism can be integrated with metacognition in the following forms:

• Social metacognition - students can collaborate with each other to reflect within group contexts. They can also share goals and can create projects as a group.
• Vygotsky’s scaffolding- AI-driven tools can be use to provide tailored guidance that complements learners' metacognitive processes.

What is Generative AI?

What is Generative AI? [Video]. InterSystems Learning Services. (2024, May 30). YouTube. https://www.youtube.com/watch?v=zMfiV0RhrCA

Generative AI (GenAI) refers to artificial intelligence systems designed to create new content, such as text, images, code, or audio, by learning patterns and generating outputs that mimic human-like creativity (OpenAI, 2023). From the video, it explains that unlike traditional AI, GenAI aims to create new data such as realistic images, text, and sounds. Trained on massive datasets, Generative AI models adeptly process and produce texts with human-like finesse. 

In the field of education, GenAI focuses on enhancing teaching and learning by generating personalized educational materials, facilitating inquiry-based learning, and promoting engagement through interactive tools. Studies have shown that tools like ChatGPT can promote metacognitive development and improve inquiry-based learning by providing immediate feedback and dynamic prompts that encourage exploration and reflection (Abdelghani et al., 2022). The most common GenAI tools in education are ChatGPT, GrammarlyGO, Quizlet Q-Chat, and Canva Magic Write. The use of GenAI aligns with modern educational goals, which explains why schools have increasingly adopted AI tools. 

GenAI is distinguished from other forms of AI by its ability to create original content and engage in open-ended interactions. Unlike broader applications in education, such as adaptive learning platforms based on fixed datasets, GenAI stands out for its ability to create dynamic and original content in response to user input. It operates as an interactive collaborator, simulating real-time engagement and creativity.

 

Role of AI in Promoting Metacognition

AI's Role in Promoting Metacognition [Video]. YouTube. (2024, June 6). https://youtu.be/cPkGTg-RLkw?si=Iihz-7fggTmU6APm

The video is one of the series developed by the team from Learn 21. The research study, Turning Research into Practice: Leveraging Generative AI in K-12 focuses on translating academic research about generative artificial intelligence (AI) into practical applications within K-12 classrooms, aiming to actively integrate AI tools to enhance student learning and teaching methods across various subjects and grade levels, while considering ethical implications and best practices for effective implementation. Furthermore, the video is designed to answer pressing questions educators and administrators face today, from harnessing AI to empower teachers to navigating its challenges, fostering deeper cognitive engagement, and promoting critical thinking and metacognition among students.

1. How to Utilize AI to Improve Metacognitive Skills:

• Encourage students to critically reflect on their interactions with AI
• Facilitate inquiry-based learning to encourage exploration and experimentation. This practice can lead to a deeper understanding and self-awareness by supporting students in engaging deeply with topics, encouraging new modes of thinking, and improving meta-linguistic knowledge (Perkins, 2023).
• Foster a collaborative environment where AI is used as a supplemental tool rather than a replacement for human interaction. By observing and discussing how others approach tasks, learners can adopt new metacognitive strategies that improve their problem-solving and self-regulation skills (Veenman, Van Hout-Wolters, & Afflerbach, 2006).Engaging with others' viewpoints can lead to more accurate self-assessment, as people often gain insights they might not achieve in isolation. This collaborative feedback loop encourages participants to think about their thinking, ultimately strengthening metacognitive skills.

2. Recommendations from the researchers in the video for effective utilization of AI tools:

• Training and ongoing coaching for educators to facilitate the effective use of generative AI. Perkins (2023) emphasizes the need for educators to be proficient in using AI tools to ensure their effective implementation in supporting metacognitive activities.
• Develop clear and consistent policies to help students use AI tools responsibly and avoid breaches of academic integrity.
• Provide continuous feedback and assessment of artificial intelligence. Su and Yang (2023) discuss the necessity of ongoing evaluation of AI tools to ensure they are meeting educational goals.
• Future research should focus on developing and evaluating AI-based educational applications and exploring their benefits and challenges in different educational contexts.

Generative AI as an Online Learning Tool

A study by Elsayary (2024) investigates how Generative AI tools can be utilized within a reflective practice model to bolster metacognitive regulation and technological proficiency of students. Reflective practice, as conceptualized by Schön (1984), is the process of learning through and from experience towards gaining new insights of self and practice. Reflective practice involves critical analysis of experiences, leading to deeper learning and is considered an essential component of metacognitive regulation (Elsayary, 2021). The study also explores the profound impact of Generative AI tools, such as ChatGPT, in the realm of school education

At a Chinese university, an AI-driven learning platform was implemented to enhance student learning experiences. This platform utilized machine learning algorithms to analyze student learning behaviors and academic performance data. Based on this analysis, it generated personalized learning materials and adaptive learning paths for each student. The platform's AI capabilities enabled it to dynamically adjust content difficulty and presentation based on individual student progress. In the said study, an AI-powered analytics tool was integrated into an online learning environment. The tool tracked and analyzed students' learning activities, providing them with visual feedback on their study habits and progress. This feedback was instrumental in helping students develop metacognitive skills, as it made them aware of their learning strategies and areas requiring improvement. Students could then adjust their study habits accordingly, leading to better academic performance.

The case study highlights the implications of these AI tools in enhancing educational engagement and effectiveness. It discusses how ChatGPT can aid in creating personalized learning experiences and foster innovative teaching. It also concludes that GenAI has the potential in revolutionizing educational practices by enhancing metacognition, active learning, and technological skills.

Generative AI can serve as a modern form of scaffolding by providing dynamic, personalized, and adaptable support that aids learners in reaching higher levels of competence. Studies and systems such as AutoTutor (Graesser et al., 2005) and Cognitive Tutor (Anderson et al., 1995) explore how AI can simulate human tutoring interactions. These systems are designed to provide scaffolding by adapting to a learner’s responses and offering contextual hints, feedback, and explanations.

AI as an Innovative Tool for Teachers

In my school district, our special education department has dipped into incorporating AI tools as a supplemental tool to help with IEP writing. We have utilized the platform called IEP Playground, which is an innovative digital platform designed to support special education by streamlining the management and development of Individualized Education Programs (IEPs). It was founded by Sean Klamm, a former special education director. The platform integrates tools such as an AI-powered "IEP CoPilot," which helps educators draft IEP goals and behavioral intervention plans more efficiently. Additionally, it includes features for real-time data access, collaboration between special and general education staff, and compliance with educational standards, all aimed at enhancing the quality of support provided to students with disabilities.

The decision to utilize AI tools such as Playground IEP was supported by the following rationale:

• Ensure consistency and conciseness of terminologies when writing an IEP,
• Reduce the time and effort needed for drafting IEP documents, allowing teachers to focus more on classroom instruction
• Enhance the quality of support provided to students with disabilities
• Foster better communication with families and promoting parental involvement in the child’s education

The platform allows educators to input and access real-time data on student performance, which can be shared with students in an age-appropriate manner. This process can encourage students to reflect on their strengths and areas for improvement, promoting self-awareness, self-regulation, and metacognition as they learn to understand their own learning processes. By providing structured supports that focus on planning, monitoring, and evaluating learning goals, Playground IEP fosters an environment where students can gradually develop metacognitive skills that are essential for lifelong learning and independence.

Use of GenAI tools in Educational Setting

In the empirical study Enhancing Metacognitive and Creativity Skills through AI-Driven Meta-Learning Strategies, the researchers applied meta-learning strategies to undergraduate students. Meta-learning is defined as a process that focuses on understanding and improving one's own learning processes. It involves using AI tools to help students identify, evaluate, and refine their learning strategies. The study employed a one-group pre-test and post-test design as its research method.

In practice, the application of meta-learning strategies led to statistically significant improvements in students' ability to plan, monitor, and reflect on their learning. One of the key implications of this research is that students can enhance their understanding of their learning strategies through meta-learning practices, such as self-reflection, goal-setting, and adaptive strategy selection. These practices, in turn, enable more effective planning, monitoring, and evaluation of one's cognitive processes, which are fundamental components of strong metacognition.

Furthermore, the use of AI tools al extends to educators' practices. A survey by Imagine Learning (2023) graphed the ways in which high school educators are currently using generative AI tools, such as ChatGPT, Bard, and DALL-E, in their classrooms:

Use of Generative AI in the Classroom [Image]. Imaginelearning. (2023, September). https://www.imaginelearning.com/wp-content/uploads/2023/12/753370622-ELE-Overview-Brochure.pdf

Based on the graph, roughly one-third of teachers surveyed currently use generative AI for the creation of assessments/tests, lesson plans, and instructional materials, as well as grading (Hallowell, 2023). Teachers are leveraging AI in high-impact areas like test creation and grading. These uses are likely seen as time-saving and productivity-enhancing. In addition, although some educators are exploring AI for customizing learning experiences, its use in student tutoring and individualized learning is still not widespread.

Students' Perceptions and Experiences with GenAI Tools

It is vital to study and understand student perceptions, as, according to Biggs (1999, 2011), student perceptions of their learning environment, abilities, and teaching strategies significantly influence their learning approach and outcomes. In line with this, the research article Students’ Voices on Generative AI: Perceptions, Benefits, and Challenges in Higher Education explores university students' perceptions of generative AI (GenAI), focusing on familiarity, willingness to engage, potential benefits, challenges, and effective integration. Overall, the participants perceived GenAI as a valuable tool with numerous benefits and expressed a willingness to use it primarily for learning, writing, and research purposes:

• Personalized and immediate support: GenAI can act as a virtual tutor, personalizing learning for students.
• Writing and brainstorming support: ChatGPT can be used as a writing assistant.
• Research and analysis support: GenAI can facilitate literature searches, summarize readings, and generate hypotheses based on data analysis.
• Visual and audio multimedia support: GenAI technologies can aid in creating artworks and handling repetitive tasks.
• Administrative support: GenAI can accelerate routine administrative tasks.

On the contrary, more than half of the participants expressed concerns about the challenges of integrating GenAI technologies, particularly regarding the reliability of the technology itself and its impact:

• Accuracy and transparency: The accuracy of GenAI responses cannot always be guaranteed, and it can be difficult to understand how AI generates answers and makes decisions.
• Privacy and ethical issues: Students are concerned about how AI might collect personal information without consent, and the issue of plagiarism is also raised.
• Holistic competencies: Over-reliance on AI may hinder the development of students' growth, skills, and intellectual abilities over time.
• Career prospects: The transformation of the workplace by GenAI has raised concerns about potential job displacement.
• Human values: There is concern that AI could misalign with human values and contribute to social injustice and inequality.
• Uncertain policies: Unclear regulations could result in the misuse or unintended consequences of GenAI.

 

In conclusion, while students acknowledge the potential of GenAI as a valuable tool for enhancing learning and productivity, they also express valid concerns regarding its reliability, ethical implications, and long-term impact on intellectual development and career prospects. Addressing these challenges through thoughtful integration, clear policies, and ongoing support will be essential for maximizing the benefits of GenAI in education.

Potentialities of Generative AI

GenAI tools, such as ChatGPT, offer opportunities for personalized learning experiences and innovative teaching methodologies, aligning with the objectives of quality education and fostering innovation.

Utilizing Generative AI has many advantages in the field of education, which the traditional education system lacks. Mittal et al., (2024) discussed the following benefits of GAI in schools:

Fostering Inquiry

A study by Abdelghani et al., (2022) highlighted how GPT-3 could become an educational tool to replicate children’s inherent curiosity and improve their ability to formulate questions. This research focused on creating automated prompts that encourage questioning, leading to more sophisticated inquiries.

Generative AI fosters inquiry by encouraging students to explore and experiment with information, generating questions, insights, and solutions through interactive engagement. It provides real-time feedback, prompts for critical thinking, and enables students to engage more deeply with content, ultimately driving curiosity and exploration in the learning process. For instance, in a study by Springer, generative AI tools like ChatGPT were found to support inquiry-based learning by offering diverse perspectives and immediate responses to student questions, prompting them to investigate topics more thoroughly and reflectively​. Hence, improving students' metacognitive knowledge and regulation. 

Inquiry requires students to reflect on what they know, what they need to know, and how they can find information. This reflection is central to metacognition, as it helps students become aware of their own understanding and areas of confusion. Research indicates that inquiry promotes metacognitive skills by encouraging students to assess their knowledge and learning strategies actively, which is essential for deeper understanding (Zohar & Dori, 2012).

Inquiry-based learning encourages students to explore questions, solve problems, and construct understanding with the support of peers, a teacher, or a technology tool. This collaborative aspect of inquiry directly corresponds to the ZPD, where a more knowledgeable "other" supports the learner in reaching a higher level of understanding (Vygotsky, 1978).

Personalized learning

Generative AI aids in personalized learning by adapting to individual students' learning needs, offering targeted feedback, and creating customized resources that align with their skill levels and interests. According to Mittal (2024), Gen AI significantly enhances personalized education by customizing content, learning pace, and complexity to meet each student’s unique requirements, learning styles, and preferences. For instance, in language learning, generative AI can correct grammar and suggest vocabulary that matches the learner’s level, fostering gradual progression in language complexity. Studies, such as those published in the International Journal of Educational Technology in Higher Education, show that tools like ChatGPT offer immediate feedback that is less biased, supporting students in self-assessment and independent learning​. In a personalized learning environment, students are encouraged to understand how they learn best, which involves recognizing which strategies and resources are most effective for them individually. This self-awareness is a fundamental component of metacognition and helps students become more self-directed learners (Flavell, 1979).

Engagement and Interactivity

It is a common observation that students often lose interest in certain classes or topics due to the subject matter or a lack of engaging teaching methods. GenAI can rekindle student interest and participation by introducing interactive and captivating educational tools, leading to improved learning outcomes (Mittal, 2024). Generative AI can automatically create educational games, and simulations, making learning more engaging and adaptive. These tools can provide students with immediate feedback, fostering active participation and quicker understanding. Additionally, teachers can use generative AI to generate discussions, problem-solving scenarios, or group activities that engage students in critical thinking. AI tools can even act as “peer” contributors in class discussions, encouraging students to articulate and defend their ideas, which strengthens their analytical skills.

Adaptability

Generative AI helps with classroom adaptability by enabling teachers to modify and personalize instruction according to students’ varied learning needs. It can adjust teaching methods according to a student’s progress and achievements, offering personalized feedback, recommendations, and learning paths. According to Diliberti (2024), the most common ways that teachers used AI tools were to adapt instructional content to fit the level of their students and to generate materials. AI is generally seen as having the potential to make teachers’ jobs easier and to support personalized instruction and special education.

Efficiency and Accessibility

According to the World Economic Forum, Generative AI can reduce the time teachers spend on repetitive administrative tasks, such as grading and creating lesson plans. For example, AI tools can automate the grading of assignments or generate instructional materials tailored to different learning objectives, allowing teachers to focus more on direct student engagement and support. It streamlines processes for teachers and makes learning materials more available to students, regardless of individual needs or challenges. For non-native speakers, generative AI can support language learning and understanding, helping students better comprehend and engage with course materials. By simplifying complex texts or offering real-time translation, generative AI reduces language barriers, making educational content more inclusive (EDUCAUSE Review).

Creativity and Innovation

Generative AI nurtures inventive thought processes and offers platforms for individual expression. Stanford educators highlight the role of generative AI in creating new kinds of learning experiences, such as collaborative projects where students can engage in metacognitive activities, enhancing their creative and critical thinking skills​. According to Adobe, generative AI fosters creativity in education by enabling students to brainstorm, generate project drafts, and explore new ideas, especially in creative fields like art and design.

In conlclusion, GenAI not only supports academic development but also nurtures metacognitive skills essential for lifelong learning. As educational systems continue to integrate GenAI, the potential for revolutionizing the learning experience becomes increasingly evident, paving the way for more inclusive, innovative, and effective educational practices.

Challenges and Limitations of AI in Educational Settings

Integrating Generative AI (GenAI) in educational settings presents several challenges. According to Stanford University, there is the issue of ensuring equitable access to these technologies, as not all institutions may have the resources to implement them effectively. The integration of GenAI also raises questions about academic integrity and the authenticity of student work, especially with tools capable of generating sophisticated content. When students rely heavily on generative AI tools, it could hinder critical thinking and engagement in developing personal insights or unique analysis, leading to a "cookie-cutter" approach to learning (Almeida et al., 2024)

Lo’s (2023) comprehensive rapid review on artificial intelligence indicates three primary limitations inherent in generative AI tools: 1. biased information, 2. constrained access to current knowledge, 3. propensity for disseminating false information. Biases in AI outputs may reflect and amplify societal stereotypes, presenting ethical and educational challenges. If students and teachers cannot identify these issues, AI could unintentionally reinforce harmful or inaccurate perspectives.

Furthermore, the integration of AI in education raises several socio-cultural concerns, including ensuring equitable access to AI tools, respecting cultural norms, maintaining personal interactions, and implementing appropriate school policies. In this context, a recent study titled Who Benefits and Who is Excluded?, explores the equity implications of using generative artificial intelligence (GenAI) in higher education. This research highlights the barriers faced by diverse student populations in accessing AI tools and how GenAI can potentially help students engage more fully in their learning.

Several countries have banned the use of ChatGPT due to concerns about misinformation, cultural misalignment, and privacy considerations. Educational institutions, such as New York City Public Schools and the Los Angeles Unified School District, have also prohibited the use of ChatGPT, believing that such AI tools lack the ability to foster critical thinking and problem-solving skills—essential components in the development of metacognition. The research also identifies other challenges, such as economic barriers, where many AI tools are subscription-based, limiting access for low-income students. Another issue is linguistic bias, which puts individuals whose primary language is not English at a disadvantage, further marginalizing these students due to insufficient support.

To address these challenges, several strategies have been proposed, including subsidizing AI tools to ensure equitable access, prioritizing diverse datasets during AI training processes to minimize bias, and providing professional development for educators to effectively integrate AI into their teaching practices. Addtionally, these limitations suggest a cautious and well-regulated approach to using AI in education, incorporating clear policies, teacher guidance, and an emphasis on balancing AI with traditional learning methods. It also justifies the reason why integrating metacognitive strategies and self-regulated learning are key to effective implementation of generative AI in education.

Empirical Studies on Critical Perspectives of Technology’s Role in Education

The study by Mittal et al. (2024), Comprehensive Review on Generative AI for Education, highlights several limitations of using Generative AI (GenAI) in educational settings:

• Copyright Issues: GenAI may generate content derived directly from existing materials, raising concerns about plagiarism and ethical boundaries.
• Unpredictable Control: Some AI models are inherently unstable, making their behavior and outputs difficult to predict or manage.
• Privacy, Security, and Ethical Concerns: GenAI has the potential to fabricate data or identities, which can facilitate impersonation and replicate copyrighted material.
• Training and Maintenance: Managing large GenAI models can be challenging for smaller educational institutions, such as schools, due to resource constraints.
• Bias, Fairness, and Ethics: GenAI may underrepresent specific racial, gender, or cultural groups. It tends to cater more effectively to tech-savvy users, potentially marginalizing others. Additionally, it can respond inconsistently to similar queries, influenced by implicit biases about the user's intent.

Interpretation of the Topic

The incorporation of AI in schools comes with both potentials and concerns. I perceive AI to be helpful for both teachers and students when utilized in a structured and correct way. I chose to study metacognition alongside self-regulated learning, sociocultural theory, and connectivism to learn what could be the effective ways to use AI in class, given the diverse needs of students. I chose this topic because many teachers do not find themselves confident or equipped with using artificial intelligence in their instruction. It can be intimidating but having a thorough understanding of the concept can be so much more beneficial.

In education, generative AI’s purposes and ideals — such as personalized learning, enhanced study skills, increase in metacognition, and self-regulated learning— can be hard to realize due to a combination of practical and infrastructural barriers. According to the U.S Department of Education, schools, especially those in under-resourced areas, often lack the infrastructure to support advanced AI tools, which require robust technology, internet access, and ongoing maintenance. The U.S. Department of Education notes that the digital divide is a critical issue, as not all schools have the technological foundation needed to implement generative AI equitably​. Additionally, research from Stanford and the Department of Education emphasizes the importance of substantial training programs to equip teachers to use AI tools effectively and responsibly. Many teachers feel underprepared, as professional development specific to AI integration is often limited. While generative AI holds promise for enhancing learning, achieving its full potential in schools requires addressing these multifaceted challenges.

Generative AI designers should consider pedagogical principles, such as goal setting, planning, self-assessment, and personalization, to ensure that AI tools effectively support students’ metacognition and improve academic performance. The following are my specific recommendations to advance the implementation of AI to increase student metacognition:

• Teachers professional development: Educators need structured training to integrate AI tools effectively into their teaching practices. Research has shown that professional development programs focusing on technology integration improve teacher confidence and student outcomes (Hinojo-Lucena et al., 2019).
• Clear and consistent policies for technology use: Educational institutions should establish transparent policies that outline ethical and practical uses of AI in classrooms. Clear guidelines help mitigate concerns about equity, data privacy, and over-reliance on technology (SpringerOpen, 2024).
• Explicit teaching of metacognition in daily activities: Incorporating metacognitive strategies, such as reflection, self-assessment, and goal-setting, into daily lessons enhances students' self-regulation (Schraw & Moshman, 1995).
• Modeling of the use of AI while teaching metacognitive skills: Teachers should demonstrate how AI can assist in tasks like planning, organizing, and monitoring progress. For example, educators can use AI-driven tools to showcase how students can set goals, track milestones, and adjust strategies based on feedback (Luckin et al., 2016).
• Use AI for Goal-Setting and Progress Tracking: Generative AI platforms can help students set personalized learning goals and track their achievements. Adaptive systems, such as Carnegie Learning and DreamBox, provide feedback on progress while suggesting actionable steps, which fosters ownership and accountability in learning.
• Facilitating Peer Collaboration through AI Platforms: AI tools can enable collaboration by assigning roles, monitoring group dynamics, and fostering equitable participation in projects. Platforms like Padlet and Miro integrate AI features that encourage brainstorming, feedback exchange, and collective reflection, aligning with social learning theories (Vygotsky, 1978).

Research Recommendations:

• AI-enhanced Curriculum: Explore strategies for incorporating adaptive learning technologies that dynamically adjust to student needs while fostering reflective learning habits.
• Long-term Impact of AI in Education: Conduct longitudinal studies to examine how sustained exposure to AI tools impacts students' academic performance, self-regulation, and lifelong learning skills.
• Fostering critical thinking skills using AI: Research the effectiveness of AI-driven simulations and scenarios in developing higher-order thinking skills

Abtahi, Y. (2018). Pupils, tools and the Zone of Proximal Development. Research in Mathematics Education, 20(1), 1-13.

Brown, A. L. (1987). Metacognition, executive control, self-regulation, and other more mysterious mechanisms. In F. E. Weinert & R. H. Kluwe (Eds.), Metacognition, motivation, and understanding (pp. 65-116). Lawrence Erlbaum Associates.

Flavell, J. H. (1979). "Metacognition and Cognitive Monitoring: A New Area of Cognitive–Developmental Inquiry." American Psychologist, 34(10), 906–911.

Graesser, A. C., Chipman, P., Haynes, B. C., & Olney, A. (2005). AutoTutor: An intelligent tutoring system with mixed-initiative dialogue. IEEE Transactions on Education, 48(4), 612–618. https://doi.org/10.1109/te.2005.856149

Hadi, M. U., tashi, qasem al, Qureshi, R., Shah, A., muneer, amgad, Irfan, M., Zafar, A., Shaikh, M. B., Akhtar, N., Wu, J., & Mirjalili, S. (2023). A Survey on Large Language Models: Applications, Challenges, Limitations, and Practical Usage. https://doi.org/10.36227/techrxiv.23589741.v1

Hinojo-Lucena, F. J., et al. (2019). Artificial Intelligence in Higher Education: A Bibliometric Study on its Impact in the Scientific Literature. Education Sciences, 9(1), 51.

Lo, C. K. (2023). What is the impact of chatgpt on education? A Rapid Review of the literature. Education Sciences, 13(4), 410. https://doi.org/10.3390/educsci13040410

Koh, E. & Doroudi, S. (2023) Learning, teaching, and assessment with generative artificial intelligence: towards a plateau of productivity, Learning: Research and Practice, 9 (2), 109-116. https://doi.org/10.1080/23735082.2023.2264086

Kremers, K. (2023). 2:02 / 2:44 Connectivism: A Networked Learning Theory for the Digital Age. YouTube. https://www.youtube.com/watch?v=ezw6_XHrwks

Luckin, R. (2018). Enhancing learning and teaching with technology: What the research says. UCL Institute of Education Press.

Norman, E. (2020). Why metacognition is not always helpful. Frontiers in Psychology, 11. https://doi.org/10.3389/fpsyg.2020.01537

Roosevelt F.D. (2008). "Zone of Proximal Development." Encyclopedia of EducationalPsychology SAGE Publication.

Saúde, S., Barros, J. P., & Almeida, I. (2024). Impacts of generative artificial intelligence in higher education: Research trends and students’ perceptions. Social Sciences, 13(8), 410. https://doi.org/10.3390/socsci13080410

Schon, D. A. (1991). The reflective practitioner. Ashgate Publishing.

Schooler, J. W., & Engstler-Schooler, T. Y. (1990). Verbal overshadowing of visual memories: Some things are better left unsaid. Cognitive Psychology, 22(1), 36–71.

Schraw, G., & Moshman, D. (1995). Metacognitive Theories. Educational Psychology Review, 7(4), 351–371.

Stanton, J. D., Sebesta, A. J., & Dunlosky, J. (2021). Fostering metacognition to support student learning and performance. CBE—Life Sciences Education, 20(2). https://doi.org/10.1187/cbe.20-12-0289

Veenman, M. V., Van Hout-Wolters, B. H., & Afflerbach, P. (2006). Metacognition and learning: Conceptual and methodological considerations. Metacognition and Learning, 1(1), 3–14.

Vygotsky, L. S. (1978). Mind in Society: The Development of Higher Psychological Processes. Harvard University Press.

Zimmerman, B. J. (2000). Attaining self-regulation: A social cognitive perspective. In M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.), Handbook of self-regulation (pp. 13-39). Academic Press.

Zimmerman, B. J. (2002). Becoming a self-regulated learner: An overview. Theory into Practice, 41(2), 64-70.

Zimmerman, B. J., & Moylan, A. R. (2009). Self-regulation: Where metacognition and motivation intersect. In D. J. Hacker, J. Dunlosky, & A. C. Graesser (Eds.), Handbook of metacognition in education (pp. 299–315).