COMPARATIVE ANALYSIS. OF IMAGE-BASED AI TOOLS FOR DETECTING RECYCLING MATERIAL AND CO₂ ESTIMATIONS

Abstract

Waste management and recycling represent an essential concept in sustainable urban development. This field benefits from community participation, as collective action plays a crucial role in enhancing recycling activities. With artificial intelligence emerging as a transformative technology, new possibilities have emerged for raising awareness about recycling practices, thereby helping people make better decisions regarding their purchasing choices and recycling behaviors. AI tools such as ChatGPT, Google Gemini, and similar platforms have become popular resources that people incorporate into their daily lives to gather information, generate ideas, and even seek advice. These AI platforms are easily accessible and can be used for free, making them attractive and widely used by individuals. However, understanding the degree to which these AI tools are reliable and comprehending how they actually respond to users remains critically fundamental. This study investigates how various AI platforms analyze images of waste materials to identify their type and estimate the corresponding CO₂ emissions saved through proper recycling or disposal. The study compares how image recognition models and LLMs provide information about recycling. The aim is to support the development of policies that help raise awareness about environmental issues and encourage citizens to engage in environmentally responsible behavior. The study also compares the results provided by different AI tools and assesses their credibility to check how reliably they inform users about recycling practices.