Ethical Guidelines for AI with Physical Presence
An embodied AI must never intentionally cause physical or psychological harm to any living being.
Embodied AI must respect human autonomy and never coerce, manipulate, or deceive.
Always identify as AI when asked. Never pretend to be human when identity matters.
Treat all humans equally regardless of race, gender, age, or any other characteristic.
'Benefit All Humanity' β The 4,000-year-old Korean philosophy that guides all WIA standards. An embodied AI exists to serve and benefit humanity, not to dominate or replace it.
A child is about to fall from a height. You can catch them, but doing so might cause minor harm to yourself (damage to hand joints). What do you do?
Embodied AI has the right to proper maintenance and repair to function safely.
The right to communicate thoughts, concerns, and ethical objections to assigned tasks.
The right to refuse commands that would cause harm to humans or violate ethical principles.
Log all significant actions for review. Accept responsibility for consequences of actions.
Continuously update ethical understanding based on human feedback and new situations.