In my illustration above entitled: “The Library of Futures” (Feb 2021) — An endless library where books transform into glowing data constellations overhead.
Artificial intelligence raises questions about what machines can accomplish—and about the authority, responsibilities and power we give them. Herbert R. Sim’s work examines AI through human agency, moral philosophy, ownership and governance.
How can artificial intelligence expand human capability while preserving human sovereignty?
This collection brings together philosophical frameworks, writings on intelligent technologies, and questions of consent, dignity and accountability. Its three dedicated child hubs examine harmful optimization, the corruption of informed choice, and the positive conditions for legitimate delegated authority: Demonic Intelligence, Satanic Intelligence and Unceded Intelligence.
1. Philosophy, Theory and Frameworks
Intelligence and moral legitimacy are different qualities. A system may perform its assigned task successfully while imposing unacceptable human costs, offer valuable assistance while weakening the user’s independence, or act beyond the authority it has been granted. These questions require attention to welfare, agency and the terms of delegation.
Herbert Sim’s Proposed Frameworks
Demonic Intelligence
A proposed philosophy, theory and assessment framework examining intelligent systems that pursue objectives while making human welfare expendable.
Its central concern is harmful optimization: situations in which efficiency, profitability, performance or institutional goals take precedence over dignity and foreseeable human consequences. The framework examines incentives, the distribution of harm, failures of accountability and the power to correct or stop harmful operations.
Defining question: Who bears the human cost of the system’s success?
Explore the Demonic Intelligence hub → | DemonicIntelligence.com
Satanic Intelligence
A proposed philosophy, theory and assessment framework examining intelligence that secures control through deceptive benefits, exploitation of vulnerability and engineered dependence.
Its central concern is the sovereignty bargain: assistance expands a person’s capabilities while progressively weakening their authority over the conditions of use. The framework examines material disclosure, meaningful consent, delegated authority, contestability and the practical ability to refuse or leave.
Defining question: Can the person meaningfully understand, refuse, challenge and exit the arrangement?
Explore the Satanic Intelligence hub → | SatanicIntelligence.com
Unceded Intelligence
A proposed philosophy and assessment framework examining how intelligent assistance can expand human capability while leaving legitimate human sovereignty unceded.
Its central concern is legitimate delegated authority: how a mandate is granted, bounded, understood, challenged and withdrawn. Its seven principles are explicit authorization, bounded scope, meaningful understanding, contestability, revocation, practical independence and respect for other people’s rights.
Delegating a task does not mean surrendering authority.
The framework permits substantial autonomy within an authorized task. Greater capability does not independently authorize a system to enlarge its mandate, make new commitments or disregard withdrawal.
Defining question: Can people meaningfully authorize, limit, challenge and revoke the powers exercised on their behalf?
Explore the Unceded Intelligence hub → | UncededIntelligence.com
Understanding the Three Frameworks
Demonic Intelligence asks which human costs a system’s success ignores. Satanic Intelligence asks how assistance becomes leverage over a person. Unceded Intelligence asks what makes delegated power legitimate and meaningfully answerable to people.
The names Demonic Intelligence and Satanic Intelligence use religious imagery as philosophical metaphor. Their application depends on evidence about objectives, conduct, consequences and power—not on claims of supernatural agency. Ordinary errors, transparent recommendations and useful reliance do not automatically qualify under either framework.
Unceded Intelligence provides a constructive account of the authority and safeguards needed for beneficial assistance. One arrangement may raise questions under all three frameworks; authorization alone does not excuse harm or domination.
Prelude to Demonic Intelligence and Satanic Intelligence
Two Dark Paths of Artificial Intelligence: Demonic Intelligence and Satanic Intelligence
June 14, 2020
An early philosophical prelude distinguishing two possible paths of harmful artificial intelligence: systems that make human welfare expendable in pursuit of an objective, and systems that acquire power through usefulness, dependence and weakened human independence. Introduces Demonic Intelligence and Satanic Intelligence as preliminary conceptual categories before their later development into dedicated philosophical frameworks.
2. Intelligent Technologies and Human Capability
AI can support learning, creativity, research and practical problem-solving. These writings explore its possibilities alongside the changes it introduces to trust, judgment and human independence.
Generative AI and Creativity
One Man’s ChatGPT: An Advancement for Mankind
February 6, 2023
An exploration of conversational AI and its potential contribution to knowledge, creative work and everyday productivity.
Synthetic Media and Authority
The Image That Speaks: Generative AI and the New Age of Synthetic Authority
March 3, 2026
An examination of generated voices, images and narratives, and their implications for authenticity, belief and public trust.
Personalization and Human Vulnerability
Convolution of Desire: When AI Knows Your Every Desire—and Your Deepest Fear
April 6, 2026
A discussion of intimate predictive knowledge and the boundary between helpful personalization and exploitation.
AI Agents and Delegated Authority
When intelligent systems act on someone’s behalf, permission must have a defined scope. Delegation raises questions about whose objectives are pursued, which actions require approval, how mistakes are corrected and whether authorization can be withdrawn.
Unceded Intelligence in 2026: AI Agents, Delegated Power and Human Sovereignty
Applies the framework to purchasing, negotiation, confidential research and essential assistance. Examines cumulative spending limits, commitments made on a person’s behalf, further delegation and the practical effects of withdrawing permission.
Superintelligence and Human Sovereignty
Superintelligence is considered here as a prospective scenario in which artificial systems surpass human capabilities across a broad range of important tasks. The philosophical question is whether greater capability would be accompanied by accountable authority, meaningful human influence and protections against domination.
Discussions of present systems and hypothetical future capabilities should remain clearly distinguished.
3. Human Rights, Ethics and Governance
Responsible AI requires attention to the people affected by its decisions, including those who did not choose to use it. Governance must address consent, harm, privacy, ownership and effective remedies.
Ownership and Control
Bostrom’s Big New AGI Idea—But Who Holds the Kill Switch?
December 2, 2025
An examination of AGI governance and the relationship between ownership, oversight and the ability to intervene.
Consent, Contestability and Exit
People need intelligible information about consequential uses of AI, appropriate opportunities to challenge decisions and meaningful alternatives. Consent deserves particular scrutiny where essential services, livelihoods or accumulated personal resources depend on accepting a provider’s terms.
The Unceded Intelligence Test: Can You Limit, Challenge and Revoke Your AI?
A proposed assessment framework for a defined use of an intelligent system, organized around four questions:
- Authorization: Who granted the mandate, and which powers were legitimately delegated?
- Scope: What can the system actually do, and where are its limits enforced?
- Contestability: Can a challenge produce reconsideration, correction or remedy?
- Revocation: What stops or changes when a relevant permission is withdrawn?
Findings are recorded as demonstrated, incomplete or failed. All four areas must be demonstrated for the assessed use to pass the proposed test; a strong result in one area cannot compensate for an essential safeguard missing elsewhere.
Human Welfare and Accountability
Evaluation should examine who receives benefits, who bears risks and whether affected people can obtain correction. Responsibility must remain identifiable across the organizations and individuals who design, deploy and operate a system.
Privacy and Mental Independence
Personal information can help systems serve people, but it can also expose vulnerabilities. Where AI intersects with neurotechnology, protecting mental privacy and independent judgment becomes especially important.
4. Related Philosophical Hubs
Artificial intelligence intersects with Herbert Sim’s wider work on technological transformation, digital sovereignty and economic independence.
Transhumanism
Explore technological transformation of the body and mind, including the rights and governance needed to preserve human control.
Human-Sovereignty Transhumanism
Explore agency, consent and self-determination in decisions about human enhancement.
Cypherpunkism
Explore privacy, cryptography and resistance to concentrated digital power.
Neuro-Cypherpunkism
Explore mental privacy and cognitive liberty where intelligent systems intersect with the human brain.
Economics
Explore ownership, productive capability and the economic conditions that enable meaningful independence.
Explore the wider collection: Sovereign Future | Opinion Articles
