is a graphic and user experience designer, researcher, educator, and amateur sewist.
DesignEd: ML-Assisted Online Course Material Preparation Tool for Design Educators, 2023
This thesis explores how the design of an online course material preparation tool that uses machine learning assist design educators in identifying their social positionality to improve accommodations for marginalized students, build empathy for intersectional out-grouped identities, diversify cultural perspectives in course materials, and encourage critical reflection to develop a more inclusive pedagogy.
User Interface, Machine Learning/AI, and User Experience
User Persona + Journey
Sophia, a new Assistant Professor of Graphic and Experience Design at NSCU, desires to adopt a more inclusive pedagogy. However, she feels overwhelmed and lacks experience interacting with diverse groups, so she seeks to use the DesignEd interface to revise her course materials to be more inclusive of marginalized students, without placing the emotional burden of education on individuals who face multiple intersections of discrimination.
Empathize
The second study proposes that the exposure to the experiences of intersectional current and former design students will strengthen empathic reactions from educators; this speculation is grounded by Keen’s theory of Narrative Empathy.
Student Narrative Archive
This narrative archive would display profiles of current and former design students (who are not presently enrolled in any of the educator’s courses) from across the country.
Since this study is situated before the semester commences, student experiential narratives will be introduced to highlight heterogeneity and incite educators’ sensitivity to and empathy for diverse student population.
Using machine learning, the system recommends profiles function suggests profiles of students who do not share much social identity overlap with user.
Profiles contain the student’s positionality information, as well as written and audio narratives.
Filters can be used to view narratives written in a specific time range. Other filtering options include sorting to show narrative media type, identity category (to find student profiles who identify as belonging to the selected marginalized category), and the search function. The search feature allows users to sift through narratives for specific words and phrases.
Diversify
The third study proposes that a design resource repository that includes resources from a range of designers and perspectives can help support educators in creating more inclusive and equitable course materials.
Design Resource Repository
By facilitating access to a wider range of design resources, educators can expose their students to a more pluriversal understanding of design and inspire them to create innovative and inclusive solutions in their own work. Ultimately, this can lead to a more just and equitable design field that celebrates and elevates the contributions of designers from all backgrounds.
.jpg)
Taking the uploaded course materials into account, ML determines the context of the course being taught and recommends resources based a few factors: relevant tags, other user’s uploaded course materials, the user’s positionality, the number of “likes” and bookmarks the resource has acquired from other users, media types the user has yet to include in their course material list (to create a more accessible resource list for students who appreciate different modalities), and data collected from the user’s behavior patterns with the system.
Dragging a resource into the ‘recommendations assistant’ drop box that will directly edit the uploaded course materials document. The system will automatically add it to the user’s course readings and resources list, citing it in APA format.
The same filters from the Student Narrative Archive are offered in the Design Resource Repository: identity category, date publication range, media type, and the search function. The media types offered in the repository provide an assortment of modality types to offer more accessible materials to students who require accommodations. Suggested topics are generated based on the content of the user’s uploaded course materials, targeting subjects that are lacking in representation
Reflect
The final study seeks to use critical self-reflection as motivation for educators to actively promote inclusive pedagogy for diverse student populations. In doing so, educators must embrace discomfort in addressing their own privileges through critical reflexivity, or employing self-reflection on oppression to gain a deeper understanding of oneself.
Reflection Feature
In addition to encouraging educators to “free-write” reflections, the system provides prompts to help them reflect on their behavior, sparking critical reflection on actions taken within the interface.
Using checkboxes that allow users to overlay their social positionality on top of their viewing history, the analysis radar chart aids educators in reflection through pointing out gaps in their exposure to marginalized identities. Viewing history, or the positionality associated with the student profiles the user has visited and bookmarked, enables educators to identify which identity perspectives they have not yet explored, and as a result, have not developed empathy towards. The user’s recommended profiles (in the Student Narrative Archive) will be updated to show more student profiles associated with marginalized identities outside this radar.
Reflection feature in context.
Prompts are generated based on user behavior patterns, and may ask the user to reflect on actions taken within the interface.
User Flow Map
User experience map outlining potential interactions and user journeys within the system.
Identify
The first study proposes that understanding social identity positionality may allow educators to better recognize potential areas of biases and critically reflect on how they accommodate out-grouped students.
Social Identity Positionality Wheel
The positionality wheel visualizes social identities on a matrix of power/privilege and marginalization/oppression within the context of the United States.
The positionality tool prompts users to disclose what identity category they belong to, as well as provides an option for users to specify what terms they use to self-describe. Then the system analyzes the user’s answers and projects their image onto the wheel of positionality to help them visualize their areas of privilege and marginalization and notice bias and growth areas.
The user uploads course materials, then natural language processing analyzes the user’s syllabi for potentially harmful or exclusionary content, making suggestions based on tone, inclusivity, and accessibility.