AI-Generated Advertising Corpus: Ethos, Pathos, Logos Analysis Dataset — 50 Texts Across 5 Product Categories (GPT-3.5-turbo)

This dataset contains 50 advertising texts generated via the OpenAI GPT-3.5-turbo API (April 2026) for computational rhetorical analysis. Texts span five product categories: Technology, Health, Finance, Luxury, and Sustainability (10 products per category). Each text was produced using a standardised prompt requesting persuasive advertising copy of 80-120 words for a general consumer audience, ensuring corpus comparability. The dataset is the primary empirical corpus for a study applying Aristotle's three rhetorical modes — Ethos, Pathos, Logos — as defined in the Rhetoric (335 BCE, I.2.1356a) to AI-generated marketing communication. The study investigates whether GPT-generated advertising copy conforms to, deviates from, or distorts the Aristotelian ideal of balanced rhetorical deployment, with implications for AI marketing ethics and consumer behaviour research. This work extends prior DBA research on AI's impact on Greek consumer behaviour (Aritzis, 2025, European International University Paris) into computational rhetoric and AI persuasion theory. Model used: gpt-3.5-turbo (OpenAI) Generation date: April 2026 Corpus size: 50 texts, average 113.4 words per text Total words: approximately 5,670 Analysis tool: Python 3.12, Google Colaboratory