[D] ICML 26 - What to do with the zero follow-up questions

A viral post on an AI research forum has exposed a critical pressure point in the academic review process for major conferences like ICML (International Conference on Machine Learning). An independent researcher, working without a supervisor, submitted a paper to ICML 2026 and entered the author rebuttal phase with above-average scores. However, three of the four assigned reviewers formally acknowledged they had follow-up questions for the authors but, as the 48-hour deadline loomed, had not actually posted any queries. This left the author in an impossible position: unable to address unspecified concerns and with no formal channel to prompt the reviewers, risking the paper's fate based on unanswered, invisible criticisms.

The incident has sparked widespread discussion among AI professionals about the fairness and logistics of tight conference deadlines, especially for independent researchers and small teams without institutional support. Critics argue that the review system, while designed for efficiency, can break down when asynchronous communication fails, potentially penalizing good work. Suggestions from the community ranged from proactively answering anticipated questions in the rebuttal to directly messaging the conference's area chair, though solutions remain ad-hoc. This case underscores the growing need for more robust and transparent digital workflows in high-stakes academic publishing, where AI research careers and funding can hinge on these outcomes.