Rédiger un PREreview

This article explores the intersection of artificial intelligence and consciousness through an integrated scientific and philosophical perspective. Advances in large language models (LLMs) and multimodal generative architectures, such as GPT-4, have intensified debates on whether artificial systems can transcend behavioral mimicry to approximate conscious awareness. Centered on the concept of imitation of the other, this study critically examines whether AI research should pursue the creation of self-aware systems or focus on developing simulations that generate convincing impressions of empathy and understanding to enhance human–AI interaction. Drawing on phenomenology and theories of alterity, the analysis argues that relational simulation—designing AI to ethically engage through simulated otherness—offers more sustainable and socially grounded benefits than the pursuit of autonomous machine consciousness. Empirical findings on simulated empathy in LLMs are integrated with current debates on AI alignment and machine ethics to define consciousness impressions as human interpretations of AI behavior. These results indicate that framing AI as a platform for relational simulation provides a practical and ethically coherent pathway for its development while serving as an experimental framework to probe the scientific and philosophical boundaries of mind, identity, and existence.