Communicating Ethical Considerations in Generative AI Systems

Abstract

Introduction: The growing adoption of Generative AI raises ethical concerns, despite gains in automation and efficiency. Objective: This study analyzes how popular generative AI systems communicate ethical considerations to users. Methodology: We applied the Semiotic Inspection Method, supported by a Semiotic Engineering-based epistemic tool, to evaluate ChatGPT, Gemini, and Claude. The analysis was guided by the principles of Beneficence, Non-Maleficence, Autonomy, Justice, and Explicability. Results: Findings reveal inconsistent and opaque ethical communication across systems. The study maps the design space of generative AI in relation to ethics and demonstrates the method’s value in advancing ethical AI research.

Ähnliche Arbeiten