Understanding GPT-3 in the Context of the eLearning Industry



eLearning as an industry is one which serves as a testing place for a lot of important tech-trends. On par with the computer/video-gaming industry, which incorporates different tech-solutions like Augmented Reality, Virtual Reality, and simulated worlds; the eLearning industry also encourages the use of technology that makes the learning experience engaging and immersive. In the light of this, one should consider the growth of an emerging technology known as GPT-3. This stands for third generation Generative Pre-trained Transformer.


GPT-3 is a part of the OpenAI project. The whole project received a lot of “press” when English newspaper “The Guardian” published an article purportedly written by GPT-3. Before getting alarmed, what needs to be understood is that the software actually wrote eight different essays, and the editorial team joined the different paragraphs together to form a full-length article. One can read about this in detail here. This blog will attempt to demystify the Open-AI project, understand GPT-3, examine some of the other AI-powered content-creation tools in the market, and look at the implications of such software for the eLearning industry as a whole.


The OpenAI Project

OpenAI is a company that is funded by many leading investors including Microsoft and Khosla Ventures. The “Charter Statement” listed on the OpenAI website provides an interesting insight into how the company intends to use its research into Artificial Intelligence to build APIs and solutions that can help different organizations add automation effectively to their workflows. GPT-3 is one of the most interesting outcomes of the work being done at OpenAI.


What is GPT-3?

GPT-3 is a language prediction model and is a natural language processing system. The quality of the output of the GPT-3 system is so high that it is difficult to actually predict if it is written by a human or an AI-program. This actually reminds us of one of Isaac Asimov’s short stories wherein a robot belonging to a novelist writes a short story, which is much more interesting than any of the novelist’s works. This alarms the novelist to a great extent and he reprograms the robot to ensure that it does not write any more stories. Though Asimov wrote this as science fiction, the speed at which AI-tech is progressing really makes one wonder where fiction ends and truth begins. In September 2020, Microsoft announced licensed “Exclusive use” of GPT-3’s underlying model, though other companies and organizations could continue to use the APIs.

The following list of applications of GPT-3 is limited but makes for interesting reading:

  • GPT-3, specifically the Codex model, is the basis for GitHub Copilot, a code completion and generation software that can be used in various code editors and IDEs

  • GPT-3 is used in certain Microsoft products to translate conventional language into formal computer code.

  • GPT-3 has been used by Andrew Mayne for AI Writer, which allows people to correspond with historical figures via email

  • GPT-3 has been used by Jason Rohrer in a retro-themed chatbot project named "Project December", which is accessible online and allows users to converse with several AIs using GPT-3 technology

  • GPT-3 was used in AI Dungeon, which generates text-based adventure games. Later it was replaced by a competing model after OpenAI changed their policy regarding generated content.


Please read the Wikipedia page for GPT-3 that offers detailed insights on what GPT-3 can do.


GPT-3 and its Relevance for the eLearning Industry

The crux of eLearning design depends on two crucial factors:

  • Content

  • Visuals

Content is handled by qualified instructional designers and technical writers who build detailed storyboards with the relevant content and image/animation references and instructions for the graphic designers to build the eLearning course. This is a long process with detailed reviews taking place at different stages of content creation, graphic design, and quality checks of the final course. Time and resource-intensive, it truly takes a fair amount of time to come up with a good, thoughtfully-designed and engaging digital learning solution.


Now just for a moment, imagine a hypothetical scenario. You have an AI-program installed on your computer. You feed in broad instructions into the program in terms of what kind of training content you want, define the target audience and learning outcomes, and suggest some visuals or animations that you would like to feature in it. The AI-powered program now comes up with a proper storyboard with all the relevant content and instructions for animation. Your senior instructional designers go through the machine-generated storyboard, make requisite changes, and forward it to the graphic design team for development. The graphic design team then gives life to the storyboard and creates a proper full-fledged eLearning course.


Now isn’t this fascinating? Look at the amount of time and effort that the company saves. This is just one of the hypothetical applications of GPT-3 in eLearning. DALL.E is another interesting innovation from OpenAI. It is a new AI system that can create realistic images and art from a description in natural language. Just imagine how GPT-3 and DALL.E2 can be utilized together to build powerful eLearning storyboards, which can be then converted into a digital learning solution using rapid course authoring tools like Articulate Storyline Rise? The possibilities are truly infinite.


Popular AI-based Content Creation Tools

Jasper, Frase, ShortlyAI, and Copysmith are just a few AI-powered content tools available in the market. Jasper is highly rated by its users and is one of the most popular AI-based content generation tools in the market.


The tool focuses on marketing copy and helps marketers create copy for display advertising and simple newsletters. Frase is a tool that is used to generate longer articles and blogs and requires users to provide a listicle-based title, the tool then generates content relevant to the topic. Read this article to learn more about such AI-based content generation tools.


The Future

At S4Carlise, we are always on the lookout for new and emerging tech and invest considerable time in exploring Machine Learning and AI-powered tools that can make our lives easier and enable a great learning experience for our esteemed clients. We truly believe that in another five years AI-powered software will play an important role in how digital learning solutions are built and deployed.

If you would like to collaborate with us on your next eLearning project, do get in touch with us at sales@s4carlisle.com.