Authors: Z Saghi, S Sabouri, S Chattopadhyay Status: Under submission, 2026
Summary
A census of 4,790 papers from three software engineering conferences (2018 to 2026), paired with free-form responses from 244 SE researchers in the ICSE 2026 FoSE survey. AI accounted for 93.1% of publication growth, but it is reorganizing SE research unevenly rather than replacing non-AI work: 92.6% of LLM-era AI papers revisit established SE problems, and 95.7% of AI papers apply or evaluate AI while 4.3% question the paradigm or develop alternatives. In the survey, 20.2% named AI-driven change as a concern about the field’s future, but only 7.0% mentioned AI when asked what the community should change. The paper argues for building cumulatively across technological waves, stronger evaluation and assurance, SE-native foundations, and studying SE in an AI-transformed world.
Links
Related
USC — Adaptive Computing Experience (ACE) Lab: written within the lab. Souti Chattopadhyay: co-author.