ICRA panel warns of paper flood

The panel's discussion highlights a field struggling with its own success. The roughly exponential growth in publications since 2017 has made it difficult for researchers to keep up, which threatens robotics' core strength as an integrative field. Kunpeng Yao's case study demonstrates this challenge, showing that a manual, year-long effort to read one subfield found only about 20 percent of papers made notable contributions.

The conversation also introduced new dimensions to the LLM debate. While tools could lower the barrier to entry for literature review, Nadia Figueroa's distinction of 'fake understanding'—where an LLM can generate structured and fluent but misleading overviews—points to a risk that goes beyond mere citation errors. This could lead to a field that produces more text but develops less understanding.

While no clean fix emerged, the panel discussions imply a shared sense that the current incentive structures may be unsustainable. Greg Dudek's vivid metaphor of building sandcastle walls as a 'tsunami' approaches acknowledges that incremental reform may be insufficient. The various proposals, from changing what is rewarded to creating new tracks for different contribution types or even reimagining peer review, all share a core goal: preserving the field's ability to connect ideas across domains before the deluge makes it impossible.