<![CDATA[The income stability and agricultural productivity of small and marginal farmers in developing countries are affected by pest infestations. Severe crop losses in India are due to increased pesticide use, limited pest-detection technologies, and restricted access to real-time advisory services. Emerging technologies such as the Internet of Things (IoT), Artificial Intelligence (AI), and Large Language Models (LLMs) offer significant opportunities to develop adaptive, farmer-centric pest management systems. This study is based on a two-component mixed method approach: (1) A large scale field study of 1000 farmers in five villages in Pendurthi Mandal, Visakhapatnam District, Andhra Pradesh, India, to assess the practice of pest control, the economic burden of pests, technology awareness and readiness for adoption of technology; and (2) A simultaneous large scale field test of a low cost AI-IoT device that includes an ESP32-CAM controller, a YOLOv8 deep learning algorithm, and a vernacular Telugu language LLM advisory engine - a new development in vernacular LLM integration tested in the field at a large scale. The survey results revealed 84% pest infestation, heavy reliance on chemical pesticides (66%), growing smartphone penetration (63%), and strong willingness to adopt (76%) when supported by government subsidies and localized AI interfaces. The field testing results verified 94% system uptime and high confidence levels of 0.87-0.94 for pest detection across four major rice pest species. This study combines findings from a survey with a concurrent field trial, confirming the efficacy of affordable pest detection and promoting sustainable agricultural practices.]]>
