Cyberbullying is a socio-technical safety challenge, and industry priorities are most clearly revealed in the inventions it patents. Treating patents as forward-looking proxies of engineering attention, we analyze 192 patent families (1983–mid-2025) using a mixed-methods approach that combines descriptive analytics, dual-coded thematic analysis, and LDA/BERTopic to map eight solution paradigms and their intended beneficiaries. We observe a sharp increase in filings from the mid-2010s, dominated by AI-centric detection and moderation pipelines, with comparatively little attention to reporting, privacy, or ethics. Only 9.4% of patent families explicitly reference minors, while other high-risk groups are largely absent, reflecting uneven beneficiary coverage. This landscape provides a novel lens on who is protected, how, and by which mechanisms, and offers actionable insights for software engineering practice, education, and policy. We further translate these findings into equity-focused requirements that highlight opportunities for more accountable and inclusive cyberbullying interventions.

@inproceedings{paperbot12034,
  title = {Cyberbullying Safety Engineering by Patent: Trends, Solution Paradigms, and Societal Impacts},
  author = {Mohamad Kassab},
  booktitle = {ICSE},
  year = {2026},
  doi = {10.1145/3786581.3786928}
}