ChatGPT is a Large Language Model (LLM) that has brought broad public awareness to the field of Artificial Intelligence. With over 100 million users, many have experienced the capabilities and boundaries of generative AI through the application. Any technological advancement, however, has both positive and negative effects. These effects appear especially acute when it involves powerful tools like LLMs. As an EdTech company focused on the application of generative AI to instructional design, we are exploring how this technology can positively impact teaching and learning while considering the ethical, instructional, and logistical boundaries.
Well, it depends. The suitability of ChatGPT, like all tools, is contextual. Sure, it has many unique affordances and use cases where it is preferable, but there are also limitations to its utility in instructional planning. The chat interface, for example, does not lend itself to the lesson planning workflow nor does the chat history format for saving content and organizing it for classes. Additionally, there is a learning curve for creating high-quality learning materials, especially as subjects become more specialized. Quality control and reliability also pose concerns. However, several prompt libraries help mitigate these challenges, such as AI for Education's Prompt Library, Microsoft's Prompts for Edu, Ethan Mollick and Lilach Mollick's More Useful Things, and Gruvy Education's Prompt Library.
Eduaide.Ai, which uses both OpenAI and Anthropic products to deliver its services, is specifically designed for the instructional context. It focuses on integrating generative AI within evidence-based teaching and learning methods (what we commonly refer to in our writings as "evidence-based instructional parameters). This makes Eduaide.Ai a niche tool tailored for educators—its use cases deliberately defined.
If a more generalized tool is needed beyond the scope of teaching, ChatGPT would be the preferred choice. However, for educators looking for an AI-assisted lesson development platform, Eduaide.Ai offers an intuitive solution built on multiple LLMs. This article aims to explore the differences between the two platforms and examine how Eduaide.Ai and ChatGPT could be used to enhance the lesson development process.
Foundation models like GPT, Claude, Gemini, and Llama are designed as generalist tools and serve as the base upon which numerous AI applications are built. Eduaide is much more specialized and seeks to constrain the chaotic "creativity" of these foundation models within the context of education—of teaching and learning.
Eduaide.Ai offers a library of over 100 generative educational resources and learning objects, such as generated texts, anchor charts, lab protocols, and IEP outlines. The Eduaide workflow, therefore, is one of integration, where all the tools needed to plan a lesson are clearly labeled and in one place. It is up to the teacher to sequence these instructional resources to fit their unique context and strengths as a teacher. In other words, Eduaide.Ai reduces the time needed to source educational materials, allowing teachers to focus on more pressing tasks like self-directed learning, personalizing materials, providing feedback, and making revisions to lessons and iterating their practice
More companies than just OpenAI offer large language models (LLMs). Anthropic, Meta, and Google have all developed their own foundation models. Eduaide, for the sake of transparency, uses models from both OpenAI and Anthropic. For example, you may interact with GPT-4o on the Feedback Bot, while the chat feature may use a specialized model enhanced by Retrieval Augmented Generation (RAG) that draws from a vast learning sciences literature base.
Additionally, Eduaide.Ai is built upon a redundancy system to seamlessly switch between models if there is a service outage from a provider. This ensures maximum up-time, even if ChatGPT or Claude is down. This model agnosticism allows us to test different models, monitor error rates, and select the best tool for the job all within one application.
Chat interfaces are helpful in many cases. However, sometimes you need to edit a generated response. Eduaide.Ai enables users to edit and transform responses right in the workspace. We even offer a suite of personalization tools to assist you in this task.
When using ChatGPT, you must know what to ask to get the desired response. The input's internal logic and context affect the output quality. This can be a timeconsuming, iterative process that is hard to replicate across the diverse resources teachers use. Eduaide.Ai removes this barrier, allowing teachers to focus on personalizing and differentiating generative resources to suit their needs without extensive training in prompt engineering and instructional design. The idea is to compliment teacher's instructional expertise with the vast computational power of LLMs.
Through techniques like RAG, prompt chaining few-shot prompting with in-context learning, model differentiation, and auto-prompting Eduaide.Ai offers a curated set of resources that integrate best practices from the Science of Learning, based on our research, collaboration with our advisory board, testing, and user feedback. Instructional planning with AI becomes an act of refinement, iteration, revision, and remixing. This generative AI workflow reduces the cost of experimentation allowing differentiated versions of assignments to be created in seconds—a lengthy text can be quickly chunked into smaller, manageable sections with clear headings, a glossary of key terms, and a group reading protocol with insightful meaning making questions.
While GPT is the popular and cost-effective LLM today, its utility may wane as the AI tool space expands. Eduaide.Ai is built such that we may swap models in and out as needed to best suit the requirements of each resource. This level of modularity enables our users to interact with a range of LLMs all in one platform. This, of course, is compared to ChatGPT where you are limited to the model used by OpenAi and thus unable to leverage other providers.
The application ChatGPT is a chatbot. While this is useful for many tasks, it is not ideal for editing lesson plans, writing student feedback, developing review games, or building differentiated assessments. With ChatGPT, you must export all AI outputs, make changes and format in a text editor or word processor, and then deploy to a Learning Management System like Blackboard or Brightspace. This process, which may involve multiple exports and imports, creates friction in using AI for instructional planning.
Eduaide.Ai addresses this issue with a built-in text editor that allows you to edit generative outputs within the web application. This enables real-time interaction with the AI as you make revisions and alterations to generated content, all within the same workspace where the LLM generates material.
Eduaide.Ai is developed by teachers in collaboration with engineers and designers, not by technologists imposing prescriptive measures that don't reflect instructional realities. We understand how teachers spend their time and the challenges that lead to burnout and attrition. This understanding allows us to continually refine the application to maximize its utility based on research, our own teaching experiences, and feedback from other teachers. While ChatGPT is a general-purpose tool, Eduaide.Ai is specifically built and supported by teachers for teachers.
Create educational content, offload time-consuming tasks to your AI teaching assistant, and never worry about "writers block" when creating teaching resources again.