fv.png

Conversational Human-AI Interactions Lab

[Blog] Dealing with Uncertainty in Academic Research

One of the most meaningful papers I have written started with a throwaway comment.

During an exploratory voice user interface study, one participant experienced what should have been a frustrating Alexa error. Instead, she chuckled and said, "It's like a silly child." That single metaphor—casual yet profound—reframed the interaction entirely. It transformed a technological failure into something human, something manageable.

This moment sparked my curiosity. I immersed myself in voice interface studies, excavating hundreds of participant quotes. A pattern emerged: people weren't just describing these systems functionally—they were building emotional scaffolding through metaphor. They spoke of AI assistants as pets needing training, children requiring patience, nosy roommates overstepping boundaries. These weren't mere linguistic flourishes. They were survival strategies—ways of making the unpredictable bearable.

When I submitted my first paper on this observation, academia pushed back. "Trivial," declared one reviewer. "Lacks scholarly merit," pronounced another. The rejection itself wasn't the blow—it was the suggestion that I'd been chasing shadows. That what felt meaningful was, in fact, meaningless.

This confronted me with what I now recognize as problem uncertainty—that destabilizing whisper asking, "Is this even a legitimate question?" Not whether you'll solve the puzzle, but whether the puzzle itself matters. In these moments of existential doubt, I return to evidence: Does this reflect authentic human behavior? Does it illuminate an unexplored territory? When both answers are yes, I continue despite the chorus of skepticism.

As we formalized our research—developing coding frameworks, expanding our dataset, grounding our approach in cognitive linguistics—methodological uncertainty emerged. We were charting unmapped territory; there's no established protocol for studying metaphors in human-computer interaction. Some reviewers found our methods too qualitative, others too interpretive. We responded by tightening analysis, clarifying claims, and fortifying everything with theoretical foundations. When methodological doubt paralyzes me, I visualize its inverse: What would constitute poor execution? Then I deliberately move in the opposite direction. Precision becomes my antidote to ambiguity.

Even with robust findings, outcome uncertainty lingered—that quiet dread that your results might be unpublishable or, worse, inconsequential. During these periods, second-order thinking becomes essential: If the immediate results disappoint, what knowledge remains valuable? Can our framework illuminate other questions? A worthwhile study generates more than conclusions—it creates new starting points.

Yet beneath all these lies the most profound uncertainty: trajectory uncertainty. What if years following this thread lead nowhere? What if this work never resonates beyond your own mind? No single mental model addresses this completely, but systems thinking helps. I've learned to measure momentum rather than outcomes. We celebrate submissions, not acceptances. Rejections aren't failures but calibration signals. If something falters, the process—not the person—needs adjustment. Research evolves. So does resilience.

This perspective—valuing process over product, reflection over reaction—has transformed my relationship with progress. Uncertainty isn't a void to be feared; it's an expanse of possibility. It contains the highest potential for divergence, for discovery, for growth. Certainty narrows; uncertainty expands. If every step forward feels guaranteed, you're likely not venturing far enough from the known.

The researchers who thrive aren't necessarily the most brilliant—they're those who don't crumble when clarity evaporates. The most innovative minds I've encountered followed unconventional paths. They've navigated institutional labyrinths without maps. They've rebuilt after mentors disappeared. They've transformed rejection into resolve. These aren't exceptions—they're exemplars of persisting when no obvious answer exists.

We romanticize research as illumination. More often, it's a practice of endurance. Of maintaining curiosity in darkness. Of holding something tenuous, something fragile, long enough for it to crystallize into meaning.

Sometimes, all you possess is a gossamer thread—a laugh, a metaphor, an elusive question—and your task is to follow it faithfully. Even when uncertainty creeps in.

Especially then.

Smit Desai1 Comment