Teaching Ethical AIML (Artificial Intelligence/Machine Learning) Playfully: Insights From a Middle School AIML Curriculum Implementation (Poster 5)

• In Event: Opportunities and Challenges in Teaching Algorithmic Justice in K–12 Computing Education

Abstract

Objectives: Curricula designed to bring AIML education to schools across the K-12 system and beyond are proliferating. With the implications of AIML transmogrified from theoretical to real in recent months and years (marked by the advent of efficacious generative AIML tools like ChatGPT and Midjourney), it is important to note that–like most professions–the majority of teachers are not AIML experts or have had ample AIML professional development opportunities. To address this gap, we advocate for a playful and accessible approach to AIML education available to teachers and learners across the curriculum. Our design hypothesis is that through playful, unplugged, authentic, and interest-driven activities, learners engage with fundamental AIML concepts while encountering ethical dilemmas inherent to AIML without having an AIML expert in the room.
Theoretical framework: Building upon an ongoing investigation (Stoiber et al., 2023) of teacher interaction with AIML curricula, our approach is based on the understanding that 1) teachers are not AIML content experts, and 2) teachers may be reticent–or unable–to engage in explicit instruction around AIML ethics or algorithmic justice (as was the case in our implementation context, a district navigating reactionary oversight). Our playful approach is informed by Zosh and colleagues (2017), who define learning through play as joyful, hands-on, iterative, meaningful, and socially interactive. Lastly, playful learning is aligned with our goals to provide equitable learning experiences, as learning through play is particularly beneficial for students typically disengaged in school (Dowd & Thomsen, 2021).
Methods & Data Sources: This paper reports on the second year implementation of a Design-Based Research inquiry which incorporated co-design with an 8th grade technology teacher. For this work, we focused on instances from the 9 implementation days when students surfaced ethical dilemmas related to AI, as well as salient student reactions to activities. Data collected includes audio and visual recordings of classroom activities, interviews with students and teachers, and student-generated artifacts, analyzed as a case study (Miles et al., 2020).
Results: While analysis is ongoing and our coding scheme still in development, observation and preliminary analysis made clear three points: 1) Playful, unplugged, authentic, and interest-driven AIML activities are widely enjoyed by students and easily facilitated by a teacher with little AIML expertise; 2) Students encountered ethical dilemmas unique to AIML while engaged in the curriculum, surfacing such dilemmas for the teacher and peers through questions, actions, and comments; and 3) Our partner teacher was not prepared to capitalize on, or productively engage with, ethical dilemmas students surfaced over the course of the curriculum.
Significance: Our results suggest teachers require more support in connecting the implications of AIML to their own and their students’ lives, while student reactions to the curriculum indicate it provides an accessible and engaging learning experience, one which takes little content expertise to facilitate. While it’s unrealistic to expect teachers to become AIML ethics experts, we propose leveraging the myriad and multifaceted expertise of teachers across the curriculum to engage students with ethics and AIML–as AIML itself is unrestricted to impacting one single domain or subject area.

Authors

• Andy Stoiber, University of Wisconsin - Madison

• YJ Kim, University of Adelaide