Assessing the Pedagogical Readiness of Large Language Models as AI Tutors in Low-Resource Contexts: A Case Study of Nepal's K-10 Curriculum

Collection of recent CS/AI papers highlighting practical and governance risks from deploying LLMs and agentic AI across education, hiring, healthcare, and organizational settings. Key findings include a large curriculum-alignment gap for AI tutors (pedagogical clarity, cultural contextualization, and two failure modes: “Expert’s Curse” and “Foundational Fallacy”); “LLM Nepotism” where AI-trusting candidates are favored in screening, creating organizational delegation and scrutiny-failure risks; and a privacy risk demonstrated by lightweight classifiers that can infer sensitive attributes from短