Reflecting on My First CHI Submission

Idea

The idea for the paper actually came from a friend’s frustration. He got into an argument with someone who disagreed with an opinion he’d formed partly with ChatGPT’s help, and he felt offended when that opinion was rejected — especially since he felt ChatGPT had “agreed” with him.

That moment stuck with me. It made me wonder: when an AI-assisted opinion gets challenged in a social setting, how do people actually negotiate that disagreement, and how do they recalibrate the trust they’ve built with the other person? That question became the seed of the study.

Study Preparation

I think of the study itself as its own product — we were building a set of scenarios and functions to support that core question, and treating the research design with the same rigor you’d bring to building a product. (I wrote more about how I approach research like this in an earlier post: Why I Study HCI.)

If you’re running an XR or VR study, or anything at that scale, it starts to feel a lot like working in industry — you have to handle latency, and use the right techniques to make sure real-time interaction is actually received properly by participants.

Writing

The writing was the hardest part. You have to clearly illustrate the interaction you’re examining and walk through your experiment step by step — there are a lot of variables involved, some analyzed together and some independently, so you have to explain precisely how each variable was measured and how the experiment was run. My guiding principle was reproducibility: I wanted anyone in the world to be able to replicate the study, which meant being as detailed as possible.

I owe a lot of gratitude here to my advisor, Yoyo Hou. He gave me invaluable advice on writing a paper that actually promotes your ideas while staying reproducible for readers. What I loved most about his approach was his instinct for leaving room — knowing when to slow down and dedicate more space to a genuinely complex idea instead of rushing through it. I spent a lot of time reworking sentence structure and pacing because of that lesson. I also want to thank my senior labmates, who — even while I was still an undergrad — spent hours giving me advice and proofreading my drafts. I’m deeply grateful to have had them on this project.

Submission, Review, and Rebuttal

This part wasn’t the hardest, but it might be the most crucial stage for actually getting published. At an HCI venue especially, methodology and findings are the most scrutinized (and debatable) parts of a paper — sometimes a contribution is subtle enough that it gets overlooked by reviewers outside your specific subfield, since HCI itself sits across CS and information science and draws reviewers from very different backgrounds. That means a lot of the work is really about communicating clearly to people who aren’t necessarily deep in your topic.

This year, the CHI rebuttal comments were solid and thoughtful — reviewers gave concrete advice on how to sharpen the writing, presentation, and literature review to better contribute to the research community. Receiving that feedback was genuinely invaluable.

Outcome

I got the decision in February, and it turned into my first publication at the conference. Presenting my research for the first time was nerve-racking — my advisor was still revising my slides fifteen minutes before my session (lol). The presentation really is the moment where you get to showcase your findings and make the case for why the work matters, so I felt a bit frustrated in the lead-up to it. But by the first day of the conference — my session was on day two — I already felt more at ease. Seeing researchers and HCI friends from all over the world sharing food, ideas, and conversations about where the field is headed gave me the confidence to dive into the harder questions during Q&A.