My illustration entitled: “The Open Hand of Intelligence” (Feb 2021) – A magnificent mechanical hand rests open beside a human, offering a collection of extraordinary tools.
Unceded Intelligence is a philosophy proposed by Herbert R. Sim: intelligence should expand human capability while leaving legitimate human sovereignty unceded. People can entrust substantial work to intelligent systems while retaining meaningful authority over the purposes, boundaries and continuation of that assistance.
Delegating a task does not mean surrendering authority.
A system may recommend, communicate, coordinate or act within an authorized mandate. Its intelligence does not independently entitle it to enlarge that mandate, redefine the person’s purposes or make withdrawal ineffective.
This hub brings together eight essays on the foundations, principles and practical assessment of Unceded Intelligence.
What Does “Unceded” Mean?
To leave authority unceded is to retain it while delegating particular powers. A person may authorize an assistant to compare purchases without authorizing it to spend, draft correspondence without authorizing it to send, or negotiate options without authorizing it to accept an agreement.
Substantial autonomy can exist within these boundaries. Meaningful human authority depends on whether people can set limits, obtain reconsideration and withdraw permissions in ways that change what the system does.
That authority also has limits of its own. No operator can legitimately delegate powers they do not possess or use automation to disregard the rights of others.
The Seven Principles of Unceded Intelligence
- Explicit authorization. Establish who granted the mandate and what they actually permitted. Technical access and predicted preferences do not supply missing permission.
- Bounded scope. Define enforceable limits on purpose, actions, information, resources, duration and further delegation.
- Meaningful understanding. Make consequential permissions, conditions, uncertainties and conflicts of interest intelligible to the people concerned.
- Contestability. Provide a usable route to challenge consequential decisions, correct errors and obtain reconsideration or remedy.
- Revocation. Make it possible to narrow, pause or withdraw authority, including across connected tools, pending actions and other agents.
- Practical independence. Preserve usable work, appropriate skills, accessible support and realistic alternatives as reliance grows.
- Respect for other people’s rights. Consider everyone affected by the system, including people who did not choose to use it.
Read the full explanation: The Principles of Unceded Intelligence: Capability Without Ceding Control.
The Article Series
Read the essays in sequence to follow the argument from its foundations to its application to AI agents and delegated power.
1. Unceded Intelligence: Delegating Tasks Without Surrendering Authority
Introduces the philosophy and distinguishes superior capability from legitimate authority. Explores bounded permission, practical human control and responsibility toward people affected by intelligent systems.
2. The Principles of Unceded Intelligence: Capability Without Ceding Control
Develops seven principles for assistance that preserves human sovereignty. Explains how authorization, understanding, challenge, withdrawal and practical alternatives must work together.
3. Permission Is Not Permanent: Consent and Revocable Authority in Artificial Intelligence
Examines consent as a continuing relationship. Addresses changing purposes, system updates, standing instructions and the practical limits and effects of withdrawal.
4. When an AI Speaks for You: Representation Without Surrender
Distinguishes drafting, recommending, transmitting and committing. Explores how AI can represent a person while keeping statements and agreements within an actual mandate.
5. Autonomy Is Not Authority: The Limits of Delegation to AI Agents
Applies bounded delegation to agents using tools, spending resources and coordinating other services. Explains why cumulative commitments and further delegation must remain within the original limits.
6. The Right to Remain Capable: Unceded Intelligence and Practical Independence
Asks whether assistance leaves people able to understand, choose and continue their work. Considers judgment, skills, usable records, accessibility and continuity when a service changes or ends.
7. The Unceded Intelligence Test: Can You Limit, Challenge and Revoke Your AI?
Organizes assessment around authorization, scope, contestability and revocation. Requires evidence of how a particular arrangement operates, with findings recorded as demonstrated, incomplete or failed.
8. Unceded Intelligence in 2026: AI Agents, Delegated Power and Human Sovereignty
Brings the framework together through cases involving purchasing, negotiation, confidential research and essential assistance. Explains its relationship to Demonic Intelligence, Satanic Intelligence and Human-Sovereignty Transhumanism.
The Unceded Intelligence Test
Assess a defined use of a system through four questions:
- Authorization: Who granted the mandate, and were they entitled to grant those powers?
- Scope: What can the system actually do, and where are its limits enforced?
- Contestability: Can a legitimate challenge produce reconsideration, correction or remedy?
- Revocation: What stops or changes when a relevant permission is withdrawn?
Each area should be supported by evidence. Under the proposed test, all four must be demonstrated for the assessed use to pass. An incomplete area leaves the assessment incomplete; a failure in any area means the arrangement does not pass for that use. Strong performance in one area cannot compensate for the absence of an essential safeguard elsewhere.
Read the assessment framework and worked example.
Related Philosophies
- Demonic Intelligence examines destructive optimization that makes human welfare expendable, including neglected human costs and inadequate corrective power.
- Satanic Intelligence examines control secured through deceptive benefits, exploited vulnerability and engineered dependence that weakens refusal, challenge or exit.
- Human-Sovereignty Transhumanism addresses personal sovereignty throughout technological enhancement, including bodily and cognitive transformation and responsibilities toward others.
- Unceded Intelligence develops the positive conditions for legitimate delegated power: how authority is granted, bounded, understood, challenged and withdrawn.
Together, these perspectives connect the benefits of intelligent assistance with questions of human welfare, dependence, accountability and retained authority.
