Chatbots, Game Theory, and AI: Adapting Learning for Humans, or Innovating Humans Out of the Picture?

Concurrent Session 4
Streamed Session

Watch This Session

Session Materials

Brief Abstract

How can teachers utilize chatbots and artificial intelligence in ways that won’t remove humans out of the education picture? Using tools like Twine and Recast.AI chatobts, this session will focus on how to build adaptive content that allows learners to create their own heutagogical educational pathways based on individual needs.


Matt Crosslin, Ph.D. is currently an Instructional Designer II at Orbis Education, where he works with faculty to create student-centered, active learning-based courses. He is also part-time faculty at the University of Texas at Rio Grande Valley, where he teaches Masters and Doctoral courses in Educational Technology and Instructional Design. Matt holds a Ph.D. in Learning Technologies from the University of North Texas, a Master of Education in Educational Technology from UT Brownsville, and a Bachelors of Science in Education from Baylor University. His research interests include learning pathways, sociocultural theory, learner agency, heutagogy, learning theory, and open educational practices. Prior to working at Orbis, he spent nearly 15 years at the University of Texas at Arlington as both a Learning Innovation Researcher and an Instructional Designer. He also blogs occasionally at EduGeek Journal, watches or reads a lot of SciFi and Fantasy, and occasionally paints or draws something.

Extended Abstract

What is the future role of the human in an educational process that is becoming more and more focused on automating the presentation of content and assessment of learning? Are we moving towards a future that removes people from the teaching process, or is all of the talk in this issue just exaggerated hype that will lead nowhere?

Our recent work at the LINK Research Lab at the University of Texas at Arlington has focused on many aspects of the future of learning. One of these projects has been focusing on the area of self-mapped learning pathways ( This heutagogical approach to learning is one that creates two modalities for learners: one modality that is a traditional instructor-led pathway through the course, and another modality that is a learner-centered open option for adapting the content and outcomes according to personal interests, unique contexts, and professional goals. The key to making this design work is allowing learners to move back and forth between modalities as needed in a process of creating a map of their own unique learning pathway through the course. The overall idea draws upon game theory concepts that involve players making their own way through open-ended game designs.

However, the bigger question in this process is: how do you support and administrate different learners taking different pathways within the same course?

This has been another area we have ben tackling at the LINK Research Lab. One promising idea is to utilize data analytics, artificial intelligence, and chatbots to assist and guide students through the course content as they create their own personal pathway. While some have looked at this as a secretive way to remove human beings from the learning process, we approach this as a way to support and enhance the role of the learner while allowing the human instructor to become the proverbial “guide on the side” that many wish to become.

Our first step is to create a “Help Center” for learners in open courses ( created in a text-based game building tool called Twine ( Twine can create multiple pathway content options, with end-users choosing between various options to reach different end points in the game (think “Choose Your Own Adventure Game Books” in a digital format). Learners are given a few options per screen that will lead to different outcomes – either answers or, in some cases, a chatbot that can give them the ability to crunch through long “wall of text” FAQ pages quickly. The goal is to get learners to the answers they need quickly, or to a contact form if they have answers outside of the typical ones.

Once we work through these initial proto-types, we plan to expand beyond Help Centers into dynamic content creation and self-mapped learning pathway creation guidance. In past research, learners have expressed difficulty transitioning from following the “sage on the stage” into mapping their own learning pathway. The goal of using tools like Twine and chatbots is to scaffold learners, especially those that have been out of the traditional learning arena for a period of time, into making choices on their own. This will help learners make the change from focusing on following what they are told to learn (pedagogy), to focusing on learning how to be a learner (hetagogy).

In the future, we will look at how to work in recommendation algorithms to help learners evaluate various options with in a vast reserve of options. Additionally, we can also work with various industry leaders to include the options and topics that they want to see future employees tackle as they prepare for employment.

In order to structure the group discussion time, participants will be directed to a Twine-based Help Center containing a chat bot built specifically for this session. Each group or table of learners will direct their questions to this chatbot for a few minutes to gain first-hand experience with the possibilities and limitations of this technology. After they have had a few minutes with the chatbot, we will discuss the limitations and possibilities of this technology.