Situated Imaginaries: Designing AI Futures with Computer Science Teaching Assistants

Paper Title

Situated Imaginaries: Designing AI Futures with Computer Science Teaching Assistants

Publication Info

• Topic area: Human–AI collaboration in education, focusing on teaching assistants' perspectives and practices.
• Keywords: AI in education, teaching assistants, human-AI collaboration, participatory design, speculative design, computing education, HCI, ethics, grading automation, instructional tools.

Background and Problem

• Problem / challenge: Limited research exists on how computer science teaching assistants (CS TAs) perceive, use, and imagine AI tools in their instructional roles. Most studies focus on students or faculty, neglecting TAs' intermediary position and growing influence in education.
• Significance: Understanding TAs' perspectives is critical as they bridge students and instructors, play a key role in course delivery, and represent future academic faculty. Their views on AI integration can shape the design of instructional tools and human–AI collaboration.
• Motivation and related work: Prior research has explored faculty perceptions of AI, automated instructional tools, and participatory design methods, but rarely considers TAs' unique intermediary roles or their situated perspectives. This paper addresses this gap by focusing on TAs' current AI practices, imagined futures, and ethical considerations.

Solution

• Proposed approach: Participatory design workshops with 131 CS TAs across two U.S. universities to explore their current uses of AI, envision future AI-enhanced tools, and reflect on ethical implications.
• Novelty:
  1. Development of a cross-institutional typology of AI use by CS TAs.
  2. Exploration of how institutional and disciplinary contexts shape TAs' visions of AI futures.
  3. Identification of ethical dilemmas and tensions in AI integration into TA workflows.
• Procedure and key techniques:
  • Five workshops lasting 75 minutes each, featuring activities such as feedback walls, collaborative design, and speculative prompts.
  • Data collection through surveys, design artifacts, audio recordings, and field notes.
  • Grounded theory analysis to identify themes and patterns in TAs' AI practices and designs.

Results

• Concrete findings:
  • 71.5% of TAs reported using AI in their instructional roles, with grading being the most frequent use (20.5% of notes). Other uses included creating teaching materials (13.6%), validating assessments (4.8%), and discussion board support (7.5%).
  • TAs imagined AI tools primarily as chatbots (54.8% of designs), targeting both TAs and students, with goals such as reducing workload and enhancing student services.
  • Ethical considerations were limited, with only 19.4% of designs addressing fairness, accountability, or transparency.
• Advantage over baselines:
  • Provides a nuanced understanding of TAs' AI practices and perspectives, which are often overlooked in existing research focused on faculty or students.
  • Highlights the interplay between institutional contexts and AI adoption, offering insights for tailored tool design.
• Experiments / evaluation:
  • Workshops conducted at two institutions with distinct demographics and TA support structures.
  • Analysis of 31 design artifacts and survey responses to triangulate findings.
• Limitations and future work:
  • Workshops emphasized near-term AI visions, limiting speculative range.
  • Study restricted to two institutions; broader research across disciplines and contexts is needed.
  • Recruitment materials may have biased the sample toward pro-AI participants.

Summary

This study investigates how computer science teaching assistants perceive and use AI tools, as well as their imagined futures for AI in education. Through participatory design workshops with 131 TAs, the authors developed a typology of AI use, explored institutional influences on AI visions, and surfaced ethical dilemmas. Findings reveal that TAs use AI for tasks like grading and material creation, envision chatbots as central to future tools, and rarely address data privacy or transparency. The study underscores the importance of centering TAs in HCI and education research, highlighting their role as knowledge workers and the need for thoughtful AI integration into instructional workflows.