A Multimodal Analysis of Human-Generated and Machine-Generated Advertisements in Pakistan

Abstract

This research aims to investigate the multimodal analysis of both Pakistani human-generated and machine-generated advertisements through visual and textual analysis. Advertisements have been defined by many scholars as how they play a huge role in shaping identity and helping people communicate their ideas. Several studies have been conducted comparing human-created ads and AI-created ads separately. However, the study used a qualitative yet comparative research design to explore the matter comprehensively. This research study chose 40 advertisements in total, comprising 20 human-generated advertisements taken from Social Media Platforms and 20 machine-generated advertisements taken from ChatGPT4 and DeepSeek. The study incorporated both English and Urdu-language-based advertisements. Furthermore, the study used Python and its libraries to analyze the linguistic and visual expressions in the advertisements. Conclusively, the findings showed that human-generated advertisements took control over machine-generated advertisements in terms of emotional resonance, cultural and social relevance. In total, the research suggested that AI had still had zero capacity to replace cultural insight and emotional connection ability.