Chatbots as a scaffold for student learning
If you’re here to interact with the chatbot for your course, please follow the pertinent link below.
- Chem 160 with Dr. Holme click here
- Chem 167 with Dr. Bonaccorsi click here
- Chem 178L with Dr. Pistolesi click here
- Chem 333L with Dr. Fernando click here
If you are here to see an example of a curated chatbot, please click on the chat icon on the bottom right of the page. This particular chatbot was curated for connecting context to chemistry for large lecture sections, and this demonstration focussed primarily on carbon capture technology.
Philosophy for chatbot development
The use of chatbots in large lectures sections addresses two key points:
- Provide tailored information and access to information and resources that students can access at any point, and
- Give the educator insight on how students are approaching and contextualizing chemistry
For interested educators
Interested in creating a chatbot? Here are some guidelines to consider in crafting the chatbot.
1. What information do you have available to map out a decision tree?
This could be commonly asked questions over your teaching experience, resources and information you want students to interact with, or areas of concerns that your students might gravitate towards. When first plotting out the decision tree, consider the types of questions that have a direct response (or multiple). Additionally, not everything has to be written altogether – a chatbot can have multiple speech bubbles in response to a prompt, so you can bring a couple of different (but related) answers if needed.
2. As you begin to map out your decision tree, consider what broad questions your students might ask.
This can determine how you start your chatbot and define how deep the chatbot may go into details. For example, you can map out guideposts in your chatbot by asking “Would you like to see the ways I could help you?” for folks who may be unsure what their questions are or what ways the chatbot can be useful to them. They can then walk through each interaction in the chatbot as if it was a guide. For those who want more freedom and flexibility, chatbots can be programmed to also take free text prompts that activate certain modules in a chatbot.
3. Beyond “yes/no” responses, what are some ways to encourage exploring or self-assessment that you might incorporate in the chatbot?
While the chatbot can mimic conversational interactions based on what you write up for the chatbot, it may not be able to answer students’ specific concerns. That’s not a deterrent! The chatbot can allow some flexibility in answering more broadly and giving students an opportunity to consider their specific situation. For example, students trying to understand if they’re on the right track for a writing prompt, the initial purpose of the chatbots created here, broad questions like:
- Does your idea include a way to talk about chemistry?
- Does your idea include a discussion about sustainability?
- Are you focused on the main topic of the semester (e.g., water footprint)?
You can add conditional responses tailored to what the students answer, that guides them to information that can address their specific needs.
4. Pick a platform that works best for your situation.
There are quite a few options in developing a chatbot and how to employ it in your classroom. For the sake of our research, we chose IBM Cloud’s Virtual Assistant as it was web and text-based. We could pull the exact interactions students had with the chatbot, and you can develop a single instance of a chatbot for free. For those who use the Microsoft suite at the institution, you may be able to use Microsoft Virtual Power Assistant.
5. Map out your decision tree, and test!