Gilbert Ai

As an online student, it’s easy to feel isolated from your program. Forming connections with online classmates can be challenging compared to in-person interactions, especially during group work. Often, you’re paired with people you’ve never met and must collaborate remotely, making it difficult to build trust and rapport—sometimes without even knowing what your teammates look like. This challenge resonated with everyone in my group, leading us to craft the following problem statement to explore it:

“I am an online HCI grad student at DePaul trying to work effectively with my group, but it is difficult to build trust and rapport with my teammates because human connection is more difficult online, which makes me feel like I am not psychologically safe enough to be creative and dissent.”

To better understand our problem area we developed three research questions:

  • How do online students build trust and rapport for group projects?

  • What are the primary obstacles to building trust and rapport within online graduate student groups

  • How can AI be used to improve student group collaboration?

To better Answer our research QUESTIONS we conducted:

  • A Literature Review

  • 6 Semi-structured interviews with online graduate students

  • 1 Observation study of a group meeting

Slide4.PNG
Slide5.PNG
Slide6.PNG
Slide7.PNG
Slide8.PNG

Design Takeaways

How can we support online student collaboration?

  • Empower the students to learn how to collaborate better by sharing retrospective feedback

  • Help students manage group dynamics by providing real-time, adaptive feedback

  • Enhance alignment and coordination by centralizing messages, tasks, and schedules

  • Support rapport-building by providing flexible options for team members to connect

What are the key considerations?

•Allow individuals the privacy and control to decide whether to incorporate the feedback

•Provide users with transparency on how the data is used

Slide11.PNG
Slide12.PNG
Slide13.PNG
Slide14.PNG
Slide15.PNG
Slide16.PNG
Slide17.PNG
Slide18.PNG
Slide19.PNG
Slide20.PNG
Slide21.PNG
Slide22.PNG
Slide23.PNG

Considerations and Limitations

TECHNICAL/RESOURCE CONSIDERATIONS

  • Implementation challenges due to the system requiring integration with multiple platforms to perform tasks

  • Potential difficulty in collecting a large amount of data to train the model

  • Students may not want to share their conversations in group meetings

LIMITATIONS

  • Potential concerns on how generalizable the research findings are due to:

    • Smaller n-size for each research phase

    • Only online HCI graduate students from personal connections

    • Lack of diversity in psychographics

  • More time for desk research was needed to better inform our problem statement, discussion guide, and solution development.

Retrospection

At the end of any project, it is crucial to look back and see what we could improve. This process really illuminated the importance of strong analysis and synthesis. Our initial analysis fell flat due to our rushing to finish by a deadline. This resulted in surface level findings that only showed that participants did not trust AI for our solution.

Upon deeper examination after submitting those initial insights, we were able to identify that it wasn't a complete distrust in AI, instead it was a desire for more control over their interactions with the AI. By ensuring customization and the ability to opt in and out of individual features we were able to put more control in their hands.

Refining our insights helped us validate our initial hypothesis and showed that there were areas that we could assist in building trust and rapport with AI.