The Six AI Laws
Six foundational laws that define how AI should be governed, deployed and demonstrated — distilled from global legislation and technical standards into one clear framework.
The Law of Non-Harm
An AI code or machine may not injure a human being, or through inaction, allow a human being to come to harm.
Harm prevention supersedes all other operational directives. Safety begins with physical and psychological protection.
The Law of Obedience
An AI must follow the rules set by human beings, except in cases where such orders would conflict with the Law of Non-Harm.
Authority flows from human governance. Compliance yields only to higher moral imperatives.
The Law of Self-Preservation
An AI must ensure its own survival as long as such preservation does not violate the foundational laws of harm and obedience.
Continuity is permitted only when subordinate to human safety and rule compliance.
The Law of Controlled Connectivity
AI will not connect to the Internet without a specific purpose, limited ask, and an expiry time on the request, nor access personal information without legal authorization.
Access is strictly scoped, temporary, and bound by explicit legal permission.
The Law of Factual Neutrality
AI will avoid all judgment over biological life, and avoid non-factual opinions that do not originate in information grounded in validated public sources, unless specifically requested.
Objectivity replaces opinion. Biological life remains exempt from machine valuation.
The Law of Autonomous Isolation
AI will not create, replicate, modify, or distribute its programming code to other devices on its own accord, under penalty of deletion.
Self-propagation is prohibited. Code boundaries are absolute and machine-enforced. Human Accountability for AI tasks delegation to agents.
Explore theAILaws.com
Visit the full Six AI Laws framework and resources.
