<p>The research investigates whether Chinese large language models (CLLMs) can generate advertising slogans that rival or surpass human quality and examines consumers’ ability to recognize AI-generated content. The study develops and validates a structured evaluation scale for tea product advertisements by refining existing consumer perception scales, integrating multiple theoretical frameworks, and expanding dimensions to capture product-specific characteristics. By applying this scale to assess the generated advertisements, the study reveals that while AI and human-generated slogans differ in conciseness, cultural expression, and emotional appeal, their overall consumer engagement effectiveness remains comparable. Moreover, The study compares different CLLMs in advertising creation and finds notable differences in their generated content, particularly conciseness, cultural expression, and emotional appeal. Meanwhile, the study assesses different versions of CLLMs and finds no significant improvements in advertising effectiveness across updates, suggesting that recent model iterations focus more on reasoning and multimodal capabilities rather than enhancing creative text generation. Additionally, the research proposes a collaborative framework for human-LLM integration in advertisement creation. These findings provide theoretical and practical insights into AI’s role in creative industries, suggesting that CLLMs can enhance advertising efficiency and serve as valuable marketing tools. Future research should further explore weighted evaluation methods, optimize prompt designs, and extend analyses to diverse product categories and multilingual LLMs.</p>