This study examines the effectiveness of generative AI, including tools such as ChatGPT, Claude, and Gemini, in green advertising, a domain that plays a critical role in promoting sustainable consumption and environmentally responsible markets. As AI becomes increasingly integrated into marketing practice, an important question arises as to whether AI-generated content can match or surpass human professionals in communicating environmentally relevant information, particularly in contexts where ethical meaning, emotional resonance, and trust are central. Drawing on processing fluency theory, this research conceptualizes advertising effectiveness as driven by the ease with which consumers process message content, distinguishing between perceptual fluency (ease of surface-level processing) and conceptual fluency (ease of meaning-based interpretation). Based on two large-scale experiments involving nearly 6,000 consumers in Japan, spanning 30 scenarios across six product categories, the findings show that the relative effectiveness of AI- versus human-created green advertisements depends on multiple contextual factors, including product characteristics, advertising design features, and consumer values. Specifically, AI-generated ads are more effective in contexts characterized by lower complexity and lower consumer involvement, such as utilitarian and lower-priced products, shorter and more structured messages, brand-owned platforms where narrative depth or emotional subtlety in green advertising is less expected, and audiences with weaker environmental concern, where perceptual fluency plays a dominant role. In contrast, human-generated ads outperform AI in contexts that require deeper interpretation and emotional and moral evaluation, including hedonic or higher-priced products, longer or more ambiguous messages, independent platforms where credibility is more salient, and audiences with stronger environmental values, where conceptual fluency becomes more important. These findings highlight that AI and human creators possess complementary strengths rather than substitutive capabilities. The study contributes to research on AI in marketing and green advertising by identifying processing fluency as a unifying mechanism that explains when AI or humans are more effective in sustainability communication. From a practical perspective, the results suggest that organizations can enhance the effectiveness of green advertising by strategically aligning the use of AI and human input with the complexity of the message and the values of the target audience, thereby improving consumer understanding of environmental attributes and supporting the broader transition toward more sustainable consumption patterns, while also informing responsible and effective communication strategies in practice.