Learning and Development with AI: Proactively Incorporating Remedies and Repairs

• In Event: AERA Roundtable Session Thursday 3:35 pm Four Seasons Ballroom 2-3
  In Roundtable Session: Advancing Assessment Fairness and Equity in Medical Education With Artificial Intelligence (Table 9)

Abstract

This project criticizes and applauses ways artificial intelligence can be used for research and evaluation. As the conference theme reminds us, innovation can do harm and lead to mistrust and thus would benefit from the concepts of remedy and repair. In the professions, we are particularly aware of the need to assess comprehensively and consider interventions which will be most beneficial with the fewest consequences. With this in mind, it is helpful to also challenge ourselves to be critical of our methodologies and to mitigate pitfalls. Like recent Artificial Intelligence (AI) work, we provide detailed examination of the accuracy of the comprehensive information gathered. Furthermore, we test ways to encourage a balanced view which maintains all perspectives. Most importantly, we also investigate which accompanied practices best provide triangulated information, insights, and aid learning and development goals.
While AI is used more and more for consolidating information gathered from learners and educators, we must be mindful to not advance efficiency at the expense of accuracy. Our past research as well as this research shows that overall AI can be useful especially at finding patterns, categorizing, and grouping sentiments; thus, performing well at condensing information in a timely manner. Indeed, a strength of AI is to sort data to identify trends which if beneficial in many areas such as customer/learner feedback, evaluation summaries, risk factor identification, and personalized support practices.1,3 4 We also learned that we need to be mindful of reproducibility, small errors, and level of accuracy of insights. Thus, part of our efforts aimed at developing more accurate results and working with other data components that can lead to more complete information.
In addition, we noted from prior research and our own work that output information can become broad and lack detail or many points of views. Thus, we examined how to reduce bias by maintaining different perspectives. We also engaged educators in insight-building techniques in an effort to add voices and create dialog to increase the conversations around different perspectives and how those can be validated. These types of activities proactively address needed repairs in trust building, pedagogical practices, and learning environments in ways that are very actionable.
Finally, we incorporate the information collected in ways that aid learning and development. This helps to minimize consequences of not using the feedback constructively and minimizes pitfalls around interpretation and action. It also keeps the focus on positive behaviors to include to maintain or repair learning environments. Using AI as a tool rather than an end product can provide a lens to view different perspectives, but more importantly to enhance personal development and address changes needed to create the best environments for all students to thrive. Despite its pitfalls, AI has the possibility of being a helpful tool towards just educational renewal.

Authors

• Kadian McIntosh, University of Arizona

• Shripad Sinari, University of Arizona