Image attributed to: Articulate content library
Team 3 will be applying what we have learned so far during our MALAT courses with a Critical Inquiry perspective as we’ve been learning in the first two weeks of the LRNT526 course. For our project, we will explore the topic of AI-enabled Tutor and Chatbot technology. Early last week we found an article published in the New York Times entitled Will chatbots teach your children? (Singer, 2024). We are already aware of competing EdTech companies in this space, including IBM Watson and NoodleFactory, and are interested to learn more about their claims, which are rooted in sound pedagogy, and which hold less promise. Overall we feel this subject holds promise as a complex, timely, and polarizing topic for critical analysis.
We will be applying our critical inquiry of AI tutors and chatbots to a MOOC (Massive Open Online Course) from Saylor Academy entitled Time and Stress Management. Since the four of us work in a variety of contexts (two in corporate learning, one in higher education, and one at an international school), we chose a general interest course which we could apply to each of our varying work environments. We had previously considered a course on Khan Academy, reading in Singer’s article (2024) that this platform was starting to pilot an AI-enabled tutor bot called Khanmigo in some of its courses (Singer, 2023). Unfortunately this technology is not yet available on all of Khan’s Academy’s courses, and when it is available it is only with a paid subscription. Since the purpose of our course is to examine the “‘real-life’ contexts” of AI technology and its “social, cultural and political” implications (Selwyn, 2010, p. 67) we felt the Saylor Academy’s open (free!) model and clean content would be a good choice. Since there is more course content (12 hours) than we would need for this project (Time and Stress Management, 2023) we have chosen to examine just the first hour of the course.
Although we are early in our research, the readings we have completed have helped us to narrow down our research focus above. We found Selwyn’s article (2010) to be especially helpful developing a critical mindset which will help us to examine whether technology always has a net-positive impact on learning (p. 68 – 69), consider the “messy realities” of learning in technology-enabled spaces (p. 69), considering both the micro- (learner) and macro-level (societal) implications of technologies (p. 70), and looking for areas of opportunity (p. 71) within our chosen AI tutor and chatbot technology. We found Singer’s two New York Times articles (2023, 2024) helpful in providing us context to what the larger society might be thinking about these technologies. Finally, Bayne’s article (2015) has helped us contextualize this application of technology from the educator’s perspective. We look forward to sharing further insights with the class as we continue to explore this topic.
As we explore this topic further, here are the questions we’ve already identified we’d like to explore while critically examining the potential application of AI tutor and chatbots to the online asynchronous course we’ve chosen:
- What are the potential risks and benefits of using AI to power chat- and tutoring bots in online learning courses?
- How do these technologies work with the iterative AI to respond to learner interactions?
- What productivity benefits might there be for over–stretched educators?
- What pedagogical benefits might there be for the learners?
- What privacy concerns need to be considered?
- What accessibility concerns need to be considered?
- Where is the technology now and where does it have the potential to go?
- How expensive are these solutions to implement? (e.g., are they equally accessible to large- and small-scale operations?)
Have you had experience using AI tutor or chatbots, either as an educator or as a learner? What was your experience like? If you haven’t experienced this yet, what hesitations do you have or what excites you? Are there any questions you think we should consider which we have not asked above? We’re excited to hear your feedback!
Andrea, Jessica, Marni, and Matt
References
Bayne, S. (2015). Teacherbot: interventions in automated teaching. Teaching in Higher Education, 20(4), 455-467. https://doi.org/10.1080/13562517.2015.1020783
Selwyn, N. (2010). Looking beyond learning: notes towards the critical study of educational technology. Journal of Computer Assisted Learning, 26, 65–73. https://doi.org/10.1111/j.1365-2729.2009.00338.x
Singer, N. (2023, June 8). New A.I. Chatbot Tutors Could Upend Student Learning. The New York Times. https://www.nytimes.com/2023/06/08/business/khan-ai-gpt-tutoring-bot.html
Singer, N. (2024, January 11). Will Chatbots Teach Your Children? The New York Times. https://www.nytimes.com/2024/01/11/technology/ai-chatbots-khan-education-tutoring.html
Time and Stress Management. (2023, May 31). Saylor Academy. https://learn.saylor.org/course/view.php?id=699
April 18, 2024 at 10:27 am
Hi, Team 3! I am so excited to read your emergent research on this topic. I am especially pleased to see you considering the issue of cost in your analysis — often this is an underexamined issue, because we make assumptions that automation always improves value. So I am really interested to see where your analysis goes.
The questions here are great. I have to say, I have generally negative experiences with chatbots, having used them primarily in customer service contexts. Given how frustrating I have found these (admittedly less sophisticated!) tools, I am a little skeptical about how to use them in education. I have been a teacher for a long time, and I find that students, even when they are asking what is at the core a question I have been asked a million times, always come at it a different way from their own contexts. I wonder about the lateral capacities of these tools, and I worry about questions being reframed for the utility of the bots.
But I would be happy to be wrong, too. 🙂 I have spent my career working to make learning more accessible to more people, and the potential here is clear. So I am genuinely excited to have my mind changed or opened by your projects this semester!
April 18, 2024 at 10:16 pm
I will refrain from just repeating Brenna’s comments above, but I absolutely share her enthusiasm to see where you are headed with your research. I think chatbots are a really interesting space to look at as they’ve got a longer history in education that pre-dates the slightly manic stooshie (this is a Scottish term) about ChatGPT.
You ask if we’ve had experience using AI tutors or chatbots, and the answer is “yes” – I was involved in the training of Teacherbot in the Bayne paper that you reference 😉 Specifically pulling data from LMS forums in historical versions of the course in order to train the bot. You might also be interested in this capstone course at my old institution here in Canada (Athabasca University) that uses an AI simulation for an internship experience for business students. Most of the students that study at AU are already working, and so this simulated coop experience makes an opportunity accessible that otherwise would not be. https://www.athabascau.ca/syllabi/admn/admn405.html
I have had mostly positive experiences with chatbots in education (though I remain always wary of the edtech salesperson in the shiny suit) and my experiences have shown me that even slightly dumb chatbots like Teacherbot can still have a productive outcome, as they can prompt useful reflection.
But both Teacherbot and the AU virtual coop examples were built for specific courses and in close collaboration with the teachers involved. At the heart of each one is careful and intentional design and a curated and bounded data-set. They are also each additional to the teacher presence, not instead-of. This is hardly the promise of scale and massive levels of efficiency gain. Which of course takes us back to the questions of cost that you have cleverly identified.
I’m also genuinely excited to see where you go with this topic!