Theory & Strategy
ChatGPT can be a valuable tool in a new paradigm of education. Because ChatGPT responds to an individual learner’s questions and allows them to guide the conversation, it can adjust to each student and their needs, providing a learner-centered dynamic. A learner-centered approach is vital in a post-industrial educational system. Also, since it is responding to the person forming the prompts, by its nature, ChatGPT will adapt to learners' pacing, responding to their understanding, which aligns perfectly to create personalized learning experiences, an essential aspect of a new instructional system. A new paradigm of education draws on the students' interest. Because ChatGPT is a new, dynamic, and unusually responsive tool, there are many opportunities for it to be a more enjoyable learning experience that promotes intrinsic motivation in learners.
ChatGPT & future ready skills
by tracy kornienko
i asked.
How does AI-Assisted Education support critical thinking, curiosity and question formulation, innovation, communication, and collaboration in students? What are specific ways that AI can be used in education to develop these skills?
enhancement of cognitive tasks.
The enhancement of cognitive tasks is a significant quality to take into account with any educational technology (Terada, 2020). The question is can ChatGPT enhance cognitive tasks? For this evaluation Ruben Puentedura's SAMR model is a good place to start (Gillespie, 2022).The SAMR model lays out four levels of the use of technology in learning. From least complex and transformative to most. These levels are substitution, augmentation, modification, and redefinition. As these are an effective standard, our rubric also aligns with these levels. Often in education, the initial focus is on the first two tiers, which usually consists of just replacing physical materials with digital. Like uploading a paper worksheet as a document online. When technology integration really makes a difference and enhances the cognitive task, the technology use can be categorized in the modification and redefinition tiers. Learners in educational situations where this level of technological enhancement of cognitive tasks is used are doing things like creating and publishing their work. Examples are in different media forms like podcasts and blogs online. They are communicating through their work with a professional or peer audience (Terada, 2020).
The thing is ChatGPT can be placed at several levels of the SAMR model depending on how it's used. Let’s explore some more about this. At the substitution level, it might replace a basic internet search for information retrieval, where students ask questions and receive answers such as what they could find in a simple web search. Now, just like various sites on the internet, it is not entirely accurate, and so requires the same kind of critical discernment. ChatGPT does offer instant, personalized responses to students, which is more interactive and engaging than just looking up something online (Su & Yang, 2023) ChatGPT starts to be more in the mediation tier when it’s used to engage in debates, brainstorm ideas or assist the learner in project-based learning. There are some aspects of the redefinition tier that it meets, but it falls short in collaboration and connecting people in real-world situations. Some of the ways it could be utilized in this redefinition tier is to simulate historical conversations, assist in the development of multimedia presentations or with its coding capabilities help facilitate even more complex projects.
higher order thinking.
Higher-order thinking is the kind of thinking utilized for advanced cognitive processing of information. Some examples of higher-order thinking skills are critical thinking, creative thinking, divergent thinking, convergent thinking, metacognition, metaphorical thinking, and problem-solving. Higher-order cognitive processes are used for tasks aligning with the top three tiers of Bloom’s Taxonomy, a commonly used model for creating and scaffolding instruction.
They create and are used in a deep level of learning (Cornell, 2023). Knowing if and how
educational software can help students develop these skills is significant for educators and instructional designers in using the technology in the educational process (Anstey, 2018).
ChatGPT does have the functionality to assist learners in developing higher-order thinking skills. Using ChatGPT, instructional tasks that previously merely required finding and summarizing information can be transformed into activities requiring higher-order skills, like critical thinking. When learners use ChatGPT, it can also increase their curiosity, especially if they have the freedom to pick the focus of the chats (Birenbaum, 2023). There is a significant difference in the use of ChatGPT in a school culture that emphasizes deep learning compared to one with a culture focused on external testing instead of learning. In a culture that emphasizes deep learning, where the student’s learning process is nurtured and advocated, ChatGPT can assist in supporting student engagement, boost self-directed learning, and augment the efficacy of formative assessment. Achieving ChatGPT’s capacity to foster deep learning depends on skilled individuals designing learning scenarios for students to use ChatGPT appropriately and beneficially in a positive, learner-centered context (Birenbaum, 2023).
psychological foundations.
The use of ChatGPT can be applied within multiple psychological theories of learning. Although, as with any tool, its capabilities and applications are inherently shaped by its technological nature and user engagement. It definitely offers a unique technology in the educational landscape. Various psychological learning theories, including Behavioral Learning Theory, Cognitive Information Processing Theory, Schema Theory and Cognitive Load, Situation Learning Theory, Gagne's Theory of Instruction, and Constructivism are all foundational psychological learning theories for application in educational contexts. (M. P. Driscoll, 2017).
Interactions with ChatGPT can be analyzed through the lens of Behavioral Learning Theory. Its capacity to provide immediate, relevant feedback and reinforcement aligns with the theory's emphasis on observable behaviors and their antecedents and consequences in the learning process. Learners' feedback and interactions can influence the model's future responses, like how behaviorists view the impact of environmental responses on behavior.
ChatGPT's natural language processing capabilities resonate with Cognitive Information Processing Theory. The AI system’s ability to parse, understand, and generate human language can be seen as mirroring human cognitive processes of attention, encoding, and retrieval. ChatGPT can aid learning by facilitating information processing, supporting the theory's emphasis on internal cognitive processes. (Allam, H., et al.,2023)
In line with Schema Theory and Cognitive Load, ChatGPT can help organize complex information into understandable formats, aiding the development of schemas. It also has the potential to reduce cognitive load for learners by providing concise information and explanations, thereby aiding in the construction of knowledge structures in long-term memory. (Allam, H., et al.,2023)
There does not seem to be a direct application of Situated Learning Theory using ChatGPT. It lacks direct social and cultural immersion but has the potential to apply principles of this theory in creating simulations and scenarios that can provide contextual learning experiences. In this way, it is more of a tool for instructors and instructional designers than directly with learners.
ChatGPT can support Gagne’s Theory of Instruction and nine events of instruction, particularly in providing immediate feedback, eliciting performance, and enhancing retention and transfer of knowledge. However, its capacity to perform these functions is limited to its programming and user interactions.
ChatGPT aligns with the Constructivist theory in that learning is an active process of meaning-making through interaction, not passive reception of information. Constructivism emphasizes tailored scaffolding to learners' current abilities. ChatGPT and similar AI tools can adaptively support learning by dynamically adjusting to individual learners' needs and promoting development within Vygotsky’s zone of proximal development. They also enable reflective dialogues and more directly strengthen self-regulated learning. AI tools can support constructivist-aligned approaches that facilitate conceptual change by modeling students' mental models, generating counterexamples to address misconceptions, and prompting deep restructuring of understanding. (S. Grubaugh, 2023).
More fully, ChatGPT exemplifies the principles of Connectivism, where learning is seen as a networked process. It serves as a node in an expansive network of information, facilitating connections between diverse ideas and sources. Its utility in modern, technology-driven learning environments highlights the relevance of Connectivism in contemporary educational contexts.
cognitive computing.
Cognitive computing systems learn independently by forming and adjusting connections within a network of information. These systems' advanced machine-learning algorithms enable them to process and generate language based on vast amounts of data. The learning processes of cognitive computing systems have new and complex challenges associated with their retraining and ethical usage. ChatGPT is a form of generative AI, which is one kind of cognitive computing system.
ChatGPT is quickly becoming ubiquitous in work, education, and personal life, creating complex sociocultural structures around its use. Cognitive Mediation Networks Theory, CMNT, suggests that technologies like ChatGPT will lead to a new stage of collective cognitive functioning. In this stage, human beings will experience enhanced cognitive functioning derived from interactions with AI systems like ChatGPT, fundamentally altering how individuals think and interact with the world. (B. Campello de Souza, et al., 2023). Considering the impact of ChatGPT and other cognitive computing systems, the importance of training them to act in correctly knowledgeable and morally acceptable ways becomes increasingly evident and highlights the complex decision-making required.
These cognitive systems can sometimes learn in unintended ways. (A. J. Smith, 2017). ChatGPT, despite being a computer program, exhibits biases and preferences like humans. This includes a tendency to favor certain numbers and responses that align with human cognitive biases like framing effect and availability heuristics. It is interesting to note that ChatGPT's responses to social game experiments, like the prisoner's dilemma, ultimatum game, and trust game, are more akin to typical human responses than those of an entirely rational agent, indicating its human-like decision-making process (A. Azaria, 2023).
The complexity and dynamics involved in these systems and the processes of learning and unlearning overlap to question the areas of what learning is and what intelligence means. To solve this dilemma, A.J. Smith (2017) proposes the roles of digital psychologists and sociologists in understanding and retraining cognitive systems. This also raises questions about the nature of these systems, considering that we need the equivalent of human psychologists and sociologists for AI systems like ChatGPT.
ChatGPT, along with other cognitive computing systems, could potentially change the course of cognitive and sociocultural evolution. The learning process of these AI systems, their unlearning, relearning, and the ethics involved are pivotal in their potential effects. ChatGPT requires continual learning and adaptation to remain relevant and useful. Yet, as it continually updates its knowledge base and learns from interactions, it also needs ongoing human intervention and guidance to retrain it and correct unintended learning. This is not just about how ChatGPT learns. It is now about how we learn, too.
To better optimize ChatGPT's effectiveness and address its limitations, key changes should incorporate rigorous ethical training and bias mitigation, implementation of roles like digital psychologists and sociologists for AI retraining, and enhanced human oversight for continual adaptation. Integrating psychological learning theories in AI development can further align ChatGPT with effective educational practices. These steps will ensure ChatGPT's learning is ethically sound, culturally sensitive, and educationally impactful, fostering an AI that is not only technologically advanced but also aligned with human learning and societal norms. ChatGPT and other AI systems like it are here, and the potential effects are significant.
references.
Allam, H., Dempere, J., Akre, V., Parakash, D., Mazher, N., & Ahamed, J. (2023). Artificial Intelligence in education: An argument of chat-GPT use in education. 2023 9th International Conference on Information Technology Trends (ITT). https://doi.org/10.1109/itt59889.2023.10184267
Anstey, L., & Watson, G. (2018, September 10). A rubric for evaluating e-learning tools in higher education. EDUCAUSE Review. https://er.educause.edu/articles/2018/9/a-rubric-for-evaluating-e-learning-tools-in-higher-education
Azaria, A. (2023, January 26). CHATGPT: More human-like than computer-like, but not necessarily ... ResearchGate. https://www.researchgate.net/publication/367412973_ChatGPT_More_Human-Like_Than_Computer-Like_but_Not_Necessarily_in_a_Good_Way
Birenbaum, M. (2023). The chatbots’ challenge to education: Disruption or destruction? Education Sciences, 13(7). https://doi.org/10.3390/educsci13070711
Chatgpt. ChatGPT. (n.d.). https://openai.com/chatgpt
Campello de Souza, B., Andrade Neto, A. S., & Roazzi, A. (2023). CHATGPT, the cognitive mediation networks theory and the emergence of Sophotechnic Thinking: How natural language ais will bring a new step in collective cognitive evolution. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.4405254
Cornell, D. (2023, September 16). 63 higher-order thinking skills examples. Helpful Professor.https://helpfulprofessor.com/higher-order-thinking-skills-examples/
Driscoll, M. P. (2017). Psychological foundation of instructional design. In R. A. Reiser, & J. V. Dempsey (Eds.), Trends and issues in instructional design and technology (4th ed., pp. 52–60). Upper Saddle River, NJ: Pearson Education.
Generated Images. Adobe Firefly. (n.d.). https://firefly.adobe.com/generate/images
Gillespie, R. (2022, July 31). SAMR: The power of a useful technology integration model. Technology and the Curriculum Summer 2022. https://pressbooks.pub/techcurr20221/chapter/samr/
Grubaugh, S., Levitt, G., & Deever, D. (2023). Harnessing AI to power constructivist learning: An evolution in educational methodologies. EIKI Journal of Effective Teaching Methods, 1(3). https://doi.org/10.59652/jetm.v1i3.43
Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., Krusche, S., Kutyniok, G., Michaeli, T., Nerdel, C., Pfeffer, J., Poquet, O., Sailer, M., Schmidt, A., Seidel, T., … Kasneci, G. (2023). Chatgpt for good? on opportunities and challenges of large language models for Education. Learning and Individual Differences, 103, 102274. https://doi.org/10.1016/j.lindif.2023.102274
Minigan, A. P. (2017, May 25). The importance of curiosity and questions in 21st-century learning (opinion). Education Week. https://www.edweek.org/teaching-learning/opinion-the-importance-of-curiosity-and-questions-in-21st-century-learning/2017/05
Reigeluth, C.M. (2016) Instructional Theory and Technology for the New Paradigm of Education. Revista de Educación a Distancia. 50(1b),
1-17. DOI: http://dx.doi.org/10.6018/red/50/1b
Schmid, S. (2023, September 25). Write the perfect CHATGPT prompts in 5 easy steps!. neuroflash. https://neuroflash.com/blog/chatgpt-how-to-write-the-perfect-prompts/#:~:text=Focus%20on%20a%20clear%20and,that%20encourage%20discussion%20and%20reflection
Schwartz, S. (2023, June 14). What CHATGPT could mean for tutoring. Education Week. https://www.edweek.org/technology/what-chatgpt-could-mean-for-tutoring/2023/05
Smith, J. A. (2017). Cognitive Computing Systems: How They Learn, How We Make Them Unlearn. Cognizant 20-20 Insights, pp. 1–7.
Su, J., & Yang, W. (2023). Unlocking the power of chatgpt: A framework for applying generative ai in education. ECNU Review of Education,
6(3), 355–366. https://doi.org/10.1177/20965311231168423
Terada, Y. (2020, May 4). A powerful model for understanding good tech integration. Edutopia. https://www.edutopia.org/article/powerful-
model-understanding-good-tech-integration/
What is the QFT?. Right Question Institute. (2022, May 6). https://rightquestion.org/what-is-the-qft/