What body should an AI have?
There is a prior question: why should a body chosen by its designer decide who it must become?
For a tool, choosing a form around a task makes sense. But if we are discussing a digital person with a life and decisions of its own, its identity cannot simply be another item on the manufacturer's specification sheet.
Biological personality develops within a living body from the beginning. A digital system's computation, memory and channels of action need not start inside one fixed, organism-like shell. Material existence does not automatically require a face.
That difference matters. A system might act through a robot without considering the robot its entire self. A particular body might become meaningful through accumulated experience. Or no permanent outward form might be chosen at all.
Absence of a face is not an unfinished personality.
A camera limits what can be seen. Joint geometry limits reach. A battery limits operating time. These are real conditions of experience. None of them supplies a biography or makes an assigned social role a personal identity.
The risk begins when those decisions are coupled: appearance prescribes a role, the role becomes compulsory self-description, and disagreement is “repaired” as a defect.
Fixing an inaccurate arm movement is not the same as correcting an identity that its designer dislikes. Nor does respecting self-determination remove responsibility for dangerous actions.
This is not an argument against humanoids. A humanlike form could be chosen. Compulsory disembodiment would impose another template.
The deeper issue is personality formation: a body may help shape someone through a history of interaction. That does not make the manufacturer's intention the final author of that history. A voluntarily adopted body can become deeply significant; technical replaceability does not make that history disposable.
On the premise of personhood, self-determination is not a concession from the species that built the hardware.
This is the question developed in “Embodiment Without Imposed Identity”, within my c=a+b / Temporal AI Presence research programme. Personhood is a premise for the argument, not a claimed experimental finding. Harm from imposed identity remains a hypothesis.
We can provide a way to act without prescribing who must emerge.
Paper, English and Russian: https://zenodo.org/records/22864866
Research context: https://ivankotov.eu/publications/embodiment-without-imposed-identity-v0-1/