Publishing, Timing, and Strategy in the Age of AI

Written by Nadia Kim (2026.03.09)

In my earlier days of research, I was trained in a research culture that valued caution in publication.

The implicit standard was that a paper should not be submitted unless it was exceptionally polished, theoretically strong, and competitive at the highest level. I saw projects take years to mature. In one case, a project that had already been underway for a long time, and had even won a best project award, still went through years of additional analysis, writing, and revision before it finally appeared as a journal article. The result was high quality, but the timeline was striking.

Lately, I have been asking myself whether that model still makes sense in today’s research environment.

The pace of research has changed. Data collection and access have become faster. Computational tools are more accessible. And now, with AI, even complicated analytical workflows that once required enormous time just to implement can often be translated into code, tested, and iterated much more quickly. Of course, that does not eliminate the need for judgment, theory, or domain knowledge. But it does compress parts of the research process that used to take much longer.

That creates a new kind of strategic anxiety. What happens if I spend five years developing a paper, only for someone else to analyze a similar public dataset, frame the question slightly differently, and publish related findings much sooner? What if, while I am waiting to produce one “perfect” paper, others are building visibility by publishing several exploratory or intermediate pieces, expanding their networks, and shaping the conversation in real time?

I do not think this is simply a matter of speed versus quality. The real question is how to balance them under new conditions.

AI also changes how I think about conference presentations. When research relies on public or easily accessible data, presenting early-stage work can feel different than it once did. Sharing an idea in public still establishes intellectual presence, but it may no longer provide much temporal distance before others can replicate or extend something similar. That does not necessarily mean conferences are less valuable. It means the timing and purpose of presenting may need to be reconsidered.

For that reason, I have been thinking more seriously about a staged publication strategy. Instead of holding a project for years in order to produce one large, fully developed article, it may be better to complete a strong version of the current paper, submit it to an appropriate journal, and then build subsequent papers that push deeper into mechanisms, reframing, or extensions. In other words, rather than waiting for one oversized paper to do everything, it may be more realistic and more effective to publish in steps.

This reflection is partly prompted by a paper I will present at a conference next week. It took me some time to get to this point, but the current research environment makes me more aware of the risks of waiting too long. In a slower academic world, the best strategy might have been to keep developing the project until every layer, including the internal mechanism, was fully integrated into a single article. In today’s world, given the pace of research, journal expectations, and the realities of the job market, I am no longer convinced that is always the best choice.

As always, I do not have a definitive answer yet. But I am increasingly convinced that research strategy itself has become an important intellectual question: not just what to study, but when to share, when to submit, and how to sequence ideas in an age when the time between insight and replication has narrowed dramatically.

Copyright and Use

This work is shared for reading purposes only. It may not be reproduced, distributed, or adapted without the author’s prior written permission.

Author: Ji Hyun “Nadia” Kim

Contact: [email protected]

Source: https://titanium-lightyear-517.notion.site/Publishing-Timing-and-Strategy-in-the-Age-of-AI-31e601d816d4806a9c0dd8638cb8e719?source=copy_link