Agency and Morality as part of Text Entry AI Assistant Personas
Andreas Komninos
TL;DR
The paper addresses the risk that text-entry AI assistants (CHATs) erode human agency and moral reasoning when used instrumentally. It argues for a value-based, agentic CHAT persona that acts as a transparent mediator, requiring user consent and offering explanatory rationale to support ethical communication in computer-mediated contexts. Through Kantian and utilitarian perspectives, it analyzes potential benefits and harms of widespread CHAT use, including threats to authenticity, linguistic diversity, and moral development, while emphasizing design guidelines to mitigate these risks. The work proposes concrete persona design principles—transparency, consent negotiation, support over prescription, and ethical guardrails—to foster responsible deployment of text-entry AI in everyday communication. The practical impact is a framework for developing CHATs that preserve autonomy and encourage critical reflection in online interactions.
Abstract
This paper discusses the need to move away from an instrumental view of text composition AI assistants under direct control of the user, towards a more agentic approach that is based on a value rationale. Based on an analysis of moral dimensions of AI assistance in computer mediated communication, the paper proposes basic guidelines for designing the agent's persona.