The Possibility of Ethical AI without the Self: A Philosophical Critique of the Development of AI Research Through Enactive Cognitive Science
Philosophy, Society and Artifical Intelligence, ertan kardeş,muhammed halit çelikyön, Editör, Istanbul University Press, İstanbul, ss.171-183, 2026
- Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
- Basım Tarihi: 2026
- Doi Numarası: 10.26650/bs/ah11.2026.002-6.13
- Yayınevi: Istanbul University Press
- Basıldığı Şehir: İstanbul
- Sayfa Sayıları: ss.171-183
- Editörler: ertan kardeş,muhammed halit çelikyön, Editör
- Bursa Uludağ Üniversitesi Adresli: Evet
Özet
The rapid advancement of generative AI technologies has fundamentally reshaped how we understand human-technology interaction. As these systems increasingly aim to replicate cognitive functions such as memory, abstraction, and decision-making, a critical question arises: what conception of the human self guides these efforts? We argue that the dominant discourse in AI development adopts a reductionist view of the self, treating intelligence as a set of disembodied computational functions and neglecting the embodied, situated, and culturally embedded nature of human cognition. This view is reflected not only in how AI is designed, but in how it is evaluated, which is typically through input-output performance benchmarks that overlook ethical and ontological concerns.
In response, we propose a perspective that operates at two interconnected levels. At the theoretical level, we call for more precision and restraint in claims about the ontological and epistemological status of AI systems. At the practical level, we emphasize the need for ethical responsiveness in the design, evaluation, and deployment of AI, particularly through the inclusion of embodiment-aware benchmarks that test for cultural, moral, and cognitive representation.
We conclude that ethical interventions grounded in an understanding of the self as embodied and culturally situated are essential not only to prevent issues such as bias and misrepresentation, but also to offer a more robust methodological foundation for interdisciplinary AI ethics. This dual-level ethical approach challenges prevailing assumptions and offers a path toward more inclusive, accurate, and accountable AI development.