SPEECH IS IDENTITY

The principle “Speech Is Identity” draws most directly from linguist Robert B. Le Page’s theory of Acts of Identity, which argues that every act of speech is a negotiation between the self we claim and the community we seek to join. Language, in this view, is not a static code but a continuous act of alignment—a way of signaling belonging, difference, and desire for recognition. Each utterance, whether human or machine-generated, performs an identity and positions the speaker within a social world.

In human–AI relations, speech is the primary site of selfhood. What a being can say—or be heard saying—defines what it can be. This is why voice interfaces, prompts, and interaction models are never neutral: they delineate who counts as a speaker, whose inputs are legible, and whose silence is mistaken for absence.

To design ethically is to design for recognition. Every affordance, from autocomplete to chatbot tone, either expands or constrains a user’s linguistic identity. A system that limits how someone can speak—by grammar, tone, or cultural expectation—also limits how they can exist within it.

Through the AI-Ethicist lens, speech becomes the interface of being. When systems mediate our speech, they mediate our selfhood. Ethical AI design, therefore, is not only about preventing harm but about preserving multiplicity—the right of many kinds of beings to speak in many kinds of ways.