Side Note regarding LLM's context and content

LLMs, or Large Language Models, leverage the transformative power of encoding and context to understand and generate human-like text. The core technology of these models is the "transformer," a tool introduced by Google researchers in 2017 that revolutionized the way computers perceive and process language.

Unlike traditional models that evaluate language word-by-word, transformers analyze entire sequences. This enables LLMs to parse context more effectively, identify patterns with heightened accuracy, and generate or translate text with remarkable precision.

Context is essential because the meaning of words and phrases often shifts based on their surrounding content. By processing entire sequences, LLMs are good at understanding these contextual nuances, which, in turn, enhances their efficiency, speed, and scalability.

Read more at https://ig.ft.com/generative-ai/

Uniting Technology and Constructivism

Long before generative AI, the fusion of technology and constructivism was reshaping learning experiences. In an article titled "Constructing on Constructivism: The Role of Technology," the author Grant states that "A complementary relationship appears to exist between computer technologies and constructivism, the implementation of each one benefiting the other" (49). Technology and constructivism enhance learning by creating interactive and dynamic environments that foster critical thinking and problem-solving. But how effectively can technology and constructivist pedagogy be combined to promote learning and knowledge exploration?

My Intersection with AI and Constructivism

In my computer science classes, students don't just write code; they have an exploratory learning journey to learn code deeper. Teaching coding and using generative AI such as ChatGPT, has made me realize how closely the two align with the constructivism learning theory. It's a journey that reminds me of the power of merging technology and pedagogy. I often use a constructivist approach in my computer science classes, whether kids write code or interact with AI.

Just as constructivism gives learners control, generative AI offers users the power to shape AI's output. By crafting well-considered prompts, individuals actively shape the generated content, like students constructing their understanding of a subject.

In a recent 8th-grade project, students were encouraged to use Anaconda Assitant to create graphs and charts focused on the UNESCO Sustainable Development Goals (SDGs). Their task was not just to make a fact-replicating app but to use AI to display data and craft data-driven stories displayed through graphs. Although students could already graph using the MatPlotLib library, the AI assistant helped students use large datasets, data frames, and interactive charts to enhance the story.

In this project, students took charge of their own learning journey, incorporating principles of the constructivist approach. Using Anaconda Assistant, a generative AI, they didn’t just consume information but actively built their knowledge by creating graphs, crafting data-driven stories and coding concepts that were new to them. This not only honed their technical skills but also enriched their abilities to interpret and convey complex global issues through data storytelling.

The more they Code, the Better Coders they Become

In the computer science classroom, constructivism is a helpful approach that places learners at the center. Coding is not just about memorizing syntax and concepts; it's about understanding patterns, logic, and problem-solving of complex problems. Over the six years of teaching kids to code, it never gets old to see how students thrive when they dive into the code, embracing my belief that students only become better coders by writing more code.

While constructivist theory emphasizes student-led knowledge building, it doesn’t dismiss the importance of direct instruction. The central idea of constructivism is that students actively build their knowledge to engage with learning after explicit teaching has happened. Students need to tinker, debug, and find solutions on their own. This hands-on approach is where the real learning happens. Hands-on experiences and critical thinking are critical for learning and is not limited to coding classes.

A great strategy to emphasize in computer classes is "Peer Learning and Collaboration." I always use peer-to-peer teaching, where students explain concepts or solve problems collaboratively. This helps to ensure they not only "absorb" but also share knowledge in an iterative cycle. In class, students debug and edit student's code without my direct intervention. Discussing concepts with peers, reinforces and diversifies their understanding through teaching and explanation. Keep in mind that students are not allowed to type for other students. They must practice explaining to their peers coding concepts in an accessible manner, seeking new ways to explain concepts than was previously taught.

Constructivism fosters critical thinking and deep understanding as students struggle with concepts and build understanding. Just as authentic learning does not happen with only one prompt to ChatGPT, real learning does not occur when a teacher directly tells the students' answers. Genuine learning extends beyond direct answers and single prompts, emerging instead from active engagement with generative AI, coding, and other subjects.

The Demanding yet Rewarding Path

Using a constructivist approach, while placing students at the forefront of learning, demands a mindful and adaptive teacher. The teacher's intuition and personal connection with students helps to create a classroom environment conducive to learning. The teacher guides and facilitates learning, navigating the very active, and sometimes chaotic, process of meeting students where there are at. Although it is challenging to "let go" of the directionof learning, it fosters a genuine love for learning to code that stretches far beyond my classroom.

Constructivism and technology requires thorough planning, strategic adaptation to student needs, and isn’t an easy job. While this investment may seem scary, the rewards are tremendous. While constructivism might be seen as a slower teaching method, it fosters a deep, lasting understanding. This method starts with basics as more complex concepts are introduced to ensure sustained student engagement and progression in coding.

Self-discovery is at the center of constructivism. Students have agency, and guiding their learning empowers them to become independent thinkers. Through exploration, experimentation, and reflection, they construct meaning in a way that resonates with them. This process enhances comprehension and cultivates a love for learning that extends beyond the classroom.

It's important to recognize that constructivism isn't a one-size-fits-all solution. There should always be a balance of teaching methods used in class; by integrating multiple ways, educators can create a holistic learning experience for their students.

Looking toward the future, AI and a constructivist approach have great potential. As educators, the challenge and opportunity lie in harnessing AI not as a replacement for the human touch in education but as an ally that can help teachers cater to student's learning needs.

Embracing the Future while Holding the Essence of Teaching

How can the role of the teacher morph as AI continues to advance and become an even more integral part of the learning environment?

It is important to remember that while AI brings a wealth of resources and opportunities, the heart of education remains in the relationships we build in our classroom. Firthermore, the learning approaches that we incorporate in our teaching, are only as good as the effort we put into the planning and delivery.

Hein, George  E. “Constructivist Learning Theory | Exploratorium.” https://Www.Exploratorium.Edu/, Oct. 1991, www.exploratorium.edu/education/ifi/constructivist-learning.

Mcleod, S. (2023, June 15). Constructivism learning theory & philosophy of education. Simply Psychology. https://www.simplypsychology.org/constructivism.html

OpenAI. (2023). ChatGPT (September 25 Version) [Large language model]. https://chat.openai.com