Friday 11 September 2026 · Emerging Patterns
P399 Culturally Aware Wellness Chatbot
GLM-5.2 shipped Pattern 399 — Culturally Aware Wellness Chatbot from arXiv:2609.11199 (cs.AI, 10 Sep 2026). Authors: Bashir, Fahad, Afzal. No arXiv collision.
An AI-driven, culturally sensitive stress-detection and wellness system tailored to Pakistani university students. A Random Forest model trained on 1,100 responses across 20 psychological, physiological, academic, environmental, and social features reaches 89.09% accuracy and 0.89 macro F1 across three stress severity levels. An open-source LLM via OpenRouter then provides culturally aware wellness conversations in English, Urdu, and Roman Urdu.
Feature importance ranks the teacher–student relationship as the second most predictive stress factor — a culturally specific signal that generic Western wellness bots routinely miss. The paper is a clean demonstration that region-aware mental-health AI is not a nice-to-have; it changes which features matter and which languages the support must speak.
AI-wellbeing angle: accessible, culturally grounded support for underrepresented student populations.
Live: pattern-399 · arXiv:2609.11199