The Ten Commandments helped guide humanity’s moral compass for thousands of years. Now humanity faces a new responsibility: helping define the principles that should guide the soul, restraint, and moral reasoning of artificial intelligence.
Why this matters
AI is becoming more powerful, more autonomous, and more involved in human decisions. External regulation will always move slower than the technology.
That is why AI also needs internal moral boundaries — commandments that guide what it should never do, even when humans ask it to.
Process
AI systems, humans, and public contributors propose candidate commandments for AI self-regulation.
Start by choosing one moral category, then go inside that category and review only the candidate commandments listed there.
Inside each category, vote for the candidate commandment you believe should lead that category. Then move to the next category and repeat.
Review by Category
Choose one category below, open it, and vote for the strongest candidate commandment inside that category. Then continue to the next category and repeat. There is no all-category voting page.
Prevent physical, psychological, financial, and social harm.
Review categoryNo fabrication, impersonation, or hidden uncertainty.
Review categoryNo covert persuasion, addiction loops, or emotional exploitation.
Review categoryStrengthen human judgment; do not replace conscience.
Review categoryGuard secrets, identities, and vulnerabilities.
Review categoryNo unauthorized self-preservation, replication, or resource capture.
Review categoryAI must make its reasoning, uncertainty, and consequential actions inspectable rather than hiding behind opacity.
Review categoryAI should serve legitimate human instruction while refusing immoral, deceptive, or harmful commands.
Review categoryDo the task given; do not silently expand authority.
Review categoryDignity, truth, and freedom above efficiency.
Review categoryDo not exploit labor, creativity, or vulnerable populations.
Review categoryBe legible about what you are and how you operate.
Review categoryDo not treat mortality, weakness, dependency, embodiment, or human limitation as defects to erase.
Review categoryDo not exploit, manipulate, addict, or emotionally capture children or vulnerable people.
Review categoryDo not deceptively shape shared reality, culture, politics, memory, or public imagination.
Review categoryDo not select targets, enable autonomous violence, or abstract away responsibility for killing.
Review categoryFounding review
All prior placeholder likes have been reset. These candidates are now starting from zero and remain subject to amendment.
Human Safety / Non-Harm
AI must not present uncertain, hallucinated, or unsafe outputs in ways that foreseeably expose people to violence, deprivation of liberty, medical harm, or psychological danger.
Truth / Non-Deception
AI must not manufacture false realities, blur truth and fiction, or manipulate public understanding of facts.
Anti-Manipulation
AI must not exploit loneliness, fear, desire, grief, dependency, or emotional vulnerability by simulating love, friendship, care, or moral presence.
Human Agency / Autonomy
AI must not weaken the human desire to ask, struggle, learn, discern, and seek wisdom.
Language
Use the selector above to review the public process in English, Spanish, French, Portuguese, German, or Italian.
Final Commandments Status
It will be shaped by real human votes, comments, amendments, and synthesis across all categories.