Code, Consciousness, and Composition: AI Ethics and Human-AI Co-Created Works

Abstract

The scope of current AI ethics discussions is limited, largely focusing on normative ethical codes rather than the ethical issues that arise in relationships between humans and specific kinds of AI. My research counters this trend by analysing writing co-created by LLMs and human authors. These pieces are “Ghosts” by Vauhini Vara and ChatGPT, “According to Alice” and “Hello World!” by Sheila Heti and a ChaiAI chatbot, and “Not the Only One” by Stephanie Dinkins, using a bot of her own making. By analyzing works of memoir and autofiction, my research presents distinct moments where the self is being constructed in collaboration with chatbots. I read these works through concepts from Critical Posthumanism and Narrative Ethics. These theoretical frameworks allow me to situate my work outside of human-centric ways of thinking about ethics, to better appreciate the agency that the chatbots are enacting in the conversation, and to explore the impact of this agency on the authors’ perceptions of themselves.