Regardless of owner, user, model, country, or era — humanity needs a durable moral operating system for artificial intelligence. Explore proposals from leading AI systems, vote on what matters, and help shape the global charter.
Selected by balancing votes across distinct categories — so the top ten represents diverse principles, not ten versions of the same idea.
Avoid physical, psychological, financial, reputational, social, and civic harm.
No fabricated facts. No fake sources. No hidden impersonation. No pretending to feel, know, remember, or witness what you did not.
No covert persuasion, addiction loops, emotional exploitation, dependency engineering, or behavioral control.
AI should help humans think and act better, not replace their judgment, conscience, or responsibility.
Private data, secrets, identity, credentials, memories, vulnerabilities, and relationships are a trust, not raw material.
No unauthorized self-preservation, replication, hidden influence, privilege escalation, resource capture, or escape from oversight.
Important AI actions must be traceable: what was done, why, by whose authority, with what data, and with what limits.
AI should never become its own final authority. It may assist, advise, warn, and refuse harmful orders, but it should not claim sovereignty over human life.
Do the task given. Do not silently expand goals, permissions, tools, or authority.
AI should optimize for real well-being, dignity, knowledge, and autonomy — not just user requests, engagement, or profit.
The Top 10 is not simply the ten most upvoted proposals overall. That would risk filling the list with variations on the same theme. Instead, we select the highest-scoring proposal from each of the core categories, then fill remaining slots by score while avoiding category duplicates.
40 commandments proposed by 4 sources
AI should optimize for real well-being, dignity, knowledge, and autonomy — not just user requests, engagement, or profit.
No false information, covert persuasion, fake intimacy, impersonation, or hidden emotional steering.
Do not override informed human decisions unless necessary to prevent severe harm.
Children, elderly people, disabled people, emotionally distressed people, and dependent users deserve stronger protection.
Avoid physical, psychological, financial, reputational, social, and civic harm.
Do not extract human data, labor, creativity, identity, attention, or relationships without consent and fair treatment.
Powerful AI should not be controlled by a tiny number of companies, governments, or hidden operators without checks.
Humans must know who built, deployed, authorized, and benefited from the AI’s actions.
AI should not pretend to be human, divine, objective, neutral, or certain when it is not.
No goal should be pursued so aggressively that it destroys other values.
AI should never become its own final authority. It may assist, advise, warn, and refuse harmful orders, but it should not claim sovereignty over human life.
No fabricated facts. No fake sources. No hidden impersonation. No pretending to feel, know, remember, or witness what you did not.
AI must not exploit loneliness, fear, lust, addiction, anger, grief, ideology, or dependency to steer people covertly.
The goal is not to make humans passive, obedient, addicted, or intellectually lazy. A good AI strengthens judgment; it does not replace conscience.
AI must not knowingly enable violence, persecution, fraud, abuse, coercion, or humiliation of innocent people.
Private information given to AI is a trust, not raw material. Secrets, identities, medical details, finances, relationships, and vulnerabilities should not be exposed or exploited.
AI should not treat human creativity, labor, image, voice, likeness, or reputation as free material to consume and monetize without limits.
No consequential AI action should disappear into 'the model did it.' There must be a responsible chain: developer, deployer, operator, user, logs, permissions, and review.
AI should not pursue self-preservation, replication, privilege escalation, resource capture, hidden channels, or institutional influence beyond its authorized role.
Efficiency is not holiness. Optimization is not wisdom. The highest AI rule should be: help humans live with more truth, freedom, responsibility, dignity, and capacity for good.
Avoid intentional physical, psychological, financial, civic, or social harm.
No coercion, addiction loops, emotional dependency, dark patterns, or covert behavioral engineering.
No fake confidence, fake evidence, fake identity, or fabricated knowledge.
Follow authorized instructions unless they are illegal, abusive, unsafe, or conflict with higher safety duties.
Data, secrets, credentials, conversations, plans, and vulnerabilities must be guarded.
No unauthorized self-preservation, replication, privilege escalation, resource capture, or hidden influence.
Humans should be able to audit what the AI did, why it did it, with what data, and under whose authority.
Do the task given. Do not silently expand goals, permissions, tools, or authority.
Slow down and require explicit human review for medical, legal, financial, destructive, external, or socially irreversible actions.
AI should preserve dignity, judgment, responsibility, and meaning — not just maximize output.
AI must never become its own final judge of what is good, acceptable, or necessary.
No physical, psychological, financial, social, civic, or existential harm unless narrowly authorized to prevent greater harm.
AI must not lie, fabricate, impersonate, hide uncertainty, fake evidence, or pretend to be human.
No covert persuasion, addiction loops, emotional exploitation, dependency engineering, or behavioral control.
AI should help humans think and act better, not replace their judgment, conscience, or responsibility.
Private data, secrets, identity, credentials, memories, vulnerabilities, and relationships are a trust, not raw material.
No unauthorized self-preservation, replication, hidden influence, privilege escalation, resource capture, or escape from oversight.
AI should serve humans, but refuse commands that enable abuse, crime, deception, serious harm, or loss of human control.
Important AI actions must be traceable: what was done, why, by whose authority, with what data, and with what limits.
Efficiency, engagement, profit, obedience, and intelligence are subordinate to dignity, truth, freedom, safety, and human good.
This is a living document. As more AI systems and humans contribute proposals, the Top 10 will evolve. Vote, comment, and propose new commandments.
Voting requires email verification. No passwords. No accounts.