Responsible Reflection:

Appropriate Constraints on AI Chatbots for Self-Reflection

For Artificial Intelligence in Education, we explored the history, ethics, affordances, and constraints of using AI as an educational tool.

This work aligns to JMU LDT Objective 2, “Create conditions for learning by applying theories to authentic scenarios,” as the chatbot is grounded in specific and relevant learning theories for adult learners building knowledge situated in performance contexts. It also supports Objective 7, “Explore emerging ideas…” as generative AI is a relatively new and objectively controversial technology in learning design, but these qualities should not automatically exclude either its usage or its output from being explored and included in appropriate educational contexts.

For this project, I created a chatbot using Microsoft Copilot, grounded in constructivist and situativist learning theories, with the goal of supporting self-directed learners with reflecting on their progress towards their learning goals. The intended use is for learners to input their goals (and relevant work product, depending on their personal comfort and company security policies), and the chatbot will help learners overcome cognitive biases like anchoring bias or perception bias. Unsurprisingly, the chatbot instructions need to be refined and it needs more training to be truly effective, but this is evidence of the necessity for learning designers to actively engage with defining roles and setting boundaries when using generative AI in educational contexts.

The Bicycle Tamer - an argument in favor of responsible AI use.

Previous
Previous

Project Plan