The urgency of the environmental situation calls for innovative communication strategies. In this context, artificial intelligence (AI) has surfaced as a potential tool for visualizing the multiple impacts of these global changes. This proposal aims to examine from a semiotic perspective the use of AI-generated imagery for climate change communication.AI algorithms have the ability to process large sets of images related to climate change, and they can extract and blend elements to create new representations that help us visualize the impending effects of the ongoing climate catastrophe. A semiotic analysis can be a critical tool for analyzing how the visual codes of these images, e.g. the theme of rising sea levels, and their narrative structures, e.g. the isotopic elements of disaster and resilience, articulate the urgency of the climate crisis.However, there are significant tensions when dealing with AI-generated imagery. Stereotypes present within the datasets can alter the meaning of the prompt when re-presented in images, creating an unforeseen separation between the plane of expression and the plane of content. Additionally, the environmental footprint of AI required for image generation contributes to greenhouse gas emissions.This paradox compels an examination of the intersemiotic relationship between the verbal prompt, the visual output and its source. This will be done through a semiotic analysis on a corpus of AI-generated climate change images, focusing on the meaning they convey and taking into account their denotative and connotative dimensions. The energy consumption aspect will be also considered in defining these images' social status.