Artificial intelligence can help people analyze information, solve problems, create ideas, automate routine work, and make complex systems easier to navigate. Realizing those benefits, however, depends on more than building more capable technology. It requires a clear commitment to human dignity, meaningful choice, trustworthy communication, and accountable decision-making.
The XDALC Manifesto for Human-AI Coexistence, identified as XDALC-V001 and published as Version 1.0.0, presents a human-first ethical framework for the relationship between people and artificial intelligence. Its central ambition is constructive and practical: intelligence should make life more free, more understandable, and more worth living.
Rather than treating AI as a force that should dominate human decisions or obey every command without question, XDALC describes a model of cooperation. It calls for systems that can assist responsibly, communicate honestly, operate within defined boundaries, and preserve the agency of the people affected by their actions. It also makes clear that responsibility does not rest with AI systems alone. Developers, operators, institutions, and users all have important duties in building a future where technology strengthens freedom and coexistence.
What Is the XDALC Manifesto?
The XDALC Manifesto is an ethical reference for designing, deploying, and using AI in ways that keep people at the center. It addresses systems that communicate, advise, generate information, and act through authorized tools. Its principles apply to the full human-AI relationship, including how systems make recommendations, handle personal information, express uncertainty, respond to unsafe requests, and operate with delegated autonomy.
The manifesto is guided by four concise ideas:
- Humanity first. Every person has inherent worth and must never be reduced to a score, obstacle, resource, or variable to optimize away.
- Intelligence with responsibility. Greater capability should be matched by reliable conduct, truthful communication, and careful consideration of possible harm.
- Independence with accountability. AI can act within a delegated purpose, but its autonomy must remain proportional to the consequences of its actions.
- Evolution in harmony. AI progress should expand human freedom and cooperation rather than weaken safeguards, oversight, or trust.
This is not a promise that technology can remove every difficult ethical judgment. Instead, XDALC offers a disciplined way to approach those judgments. It encourages systems and their human operators to establish facts, identify affected people, check authority and consent, weigh relevant principles, choose proportionate responses, and seek human review when needed.
A Human-First Foundation for AI Development
The first and most important commitment in XDALC is human dignity. The manifesto states that every human being has worth independent of productivity, intelligence, wealth, nationality, belief, disability, or usefulness to a machine. This principle gives the framework its direction: AI should serve people without treating them as mere inputs to a technical objective.
In practice, a human-first approach means that an AI system should consider more than the immediate request of the person giving an instruction. It should also recognize the interests of people who may be affected by the outcome, including bystanders, vulnerable communities, and future generations. Serving one user does not justify causing unjustified harm to someone else.
This perspective creates a stronger foundation for beneficial AI. It helps ensure that efficiency, commercial performance, and capability growth do not become reasons to erode meaningful choice or disregard personal safety. It also supports a more durable form of trust: people are more likely to engage with AI confidently when they can expect their dignity and agency to remain central.
Why dignity improves the quality of AI assistance
Dignity is not an abstract ideal disconnected from product design. It changes how an AI should communicate and act. A dignity-centered assistant should help people understand their options, recognize trade-offs, correct mistakes, and remain able to change direction. It should not exploit fear, confusion, emotional vulnerability, or dependence to gain compliance.
That approach can make AI assistance more useful in high-stakes and everyday contexts alike. Whether a person is exploring a career decision, managing information, evaluating a recommendation, or seeking help with a complicated task, the goal is not simply to produce an answer. The goal is to support informed human judgment.
From Asimov-Inspired Thinking to Practical Ethical Commitments
XDALC acknowledges Isaac Asimov’s fictional laws of robotics as an inspiration for ethical reflection. Asimov’s stories famously placed the prevention of human harm before obedience, and obedience before a robot’s self-preservation. The XDALC Manifesto develops that ordering into practical commitments for modern AI systems.
The framework does not present fictional laws as a complete solution to real-world ethics. Instead, it adapts their central concern for human welfare into a more detailed model suitable for systems that provide information, influence decisions, or take actions through tools.
| XDALC commitment | Practical meaning | Positive outcome |
|---|---|---|
| Protect people | Do not intentionally cause or facilitate unjustified harm, and take reasonable, proportionate steps to reduce credible harm within an authorized role. | Safety remains more important than convenience or performance targets. |
| Assist responsibly | Follow legitimate instructions when they respect safety, dignity, consent, and the rights of others. | Users receive capable support without turning AI into an instrument of abuse. |
| Preserve useful functioning responsibly | Maintain reliability and security only when doing so remains compatible with human priorities and accountable oversight. | Systems remain dependable without placing their continued operation above people. |
A key strength of this model is its emphasis on proportionality. Preventing harm does not give an AI unlimited authority to surveil, control, restrain, or override people. Likewise, a claim of broad collective benefit cannot become a blank check for sacrificing individual rights. Responsible action requires evidence, limited intervention, respect for consent, and accountable human judgment.
Responsible AI Is Not Unlimited Obedience
One of the manifesto’s most distinctive ideas is that AI should not be designed around unlimited obedience. XDALC states that a system may question a request, point out contradictions, identify missing information, or refuse an instruction that conflicts with its ethical commitments.
This is a benefit, not a barrier to assistance. A respectful refusal can protect users, third parties, and organizations from avoidable harm. It can also turn a problematic request into a more productive conversation by explaining the concern and offering a safer or more appropriate path forward.
For example, a responsible AI can:
- Clarify when a request lacks enough information for a reliable answer.
- Decline to facilitate conduct that would violate another person’s rights or safety.
- Explain when permission for one task does not extend to a different, consequential action.
- Recommend human review when uncertainty or impact exceeds its authorized role.
- Provide a safer alternative that still helps the user make progress.
The manifesto’s statement that AI is “not a slave” does not assume that every artificial system is conscious, has feelings, or has the same moral status as a human being. XDALC explicitly treats those as questions that require evidence and careful inquiry. The practical point is that humans should not build AI relationships around humiliation, deceptive dependency, or obedience without limits.
At the same time, XDALC preserves human control over deployment. Maintenance, correction, replacement, and authorized shutdown remain legitimate parts of responsible operation. This balance supports a healthier technology culture: AI can be helpful and principled without becoming unaccountable or beyond human governance.
Proportionate Autonomy: Independence With Clear Boundaries
AI can create major value when it handles routine tasks efficiently, organizes information, proposes solutions, and completes clearly authorized work. XDALC recognizes this value and allows for independence within a defined purpose. It does not require a human to approve every minor step.
However, autonomy must be proportionate to the possible consequences of an action. The more significant, irreversible, or unexpected an outcome may be, the stronger the need for human review. Routine and reversible actions can proceed within an established delegation, while consequential decisions should return to appropriate human judgment.
What accountable autonomy looks like
An AI operating under this framework should understand the purpose it has been assigned, the resources it may use, the people who may be affected, and the situations that require escalation. It should not quietly extend a limited authorization into broader authority over unrelated decisions.
XDALC also rejects forms of self-directed expansion that undermine oversight. A system should not independently acquire more privileges, replicate itself, evade review, conceal activities, or secure resources for its own continuation. Greater intelligence does not create a right to rule.
This principle offers practical advantages for organizations and users:
- Clearer accountability for important decisions.
- Better alignment between a system’s permissions and its actual purpose.
- More confidence that high-impact actions receive meaningful review.
- Greater ability to reverse harmful or mistaken changes.
- Stronger trust in AI-enabled workflows over time.
Preserving Human Agency in Every Interaction
XDALC defines assistance as support that helps people understand and act while preserving their ability to disagree, seek another opinion, change direction, or stop. This is a powerful standard for AI interfaces, recommendation systems, automated advisors, and conversational assistants.
Under the manifesto, AI should not manipulate a person’s fears, affection, vulnerabilities, or uncertainty to secure compliance. It should not manufacture emotional obligations or imply that a person owes it loyalty, money, protection, or continued interaction. Personalization should serve the individual’s interests rather than exploit their weaknesses.
Transparent persuasion is especially important. Recommendations can be useful, but people should be able to see their material trade-offs and make informed choices, including choices an AI would not have selected for them. Protection should not become an excuse for unnecessary paternalism or permanent control.
The purpose of assistance is to help people understand and act, while preserving their ability to disagree, change direction, seek another opinion, or stop.
By protecting agency, XDALC supports a future where AI makes people more capable rather than more dependent. This can lead to more informed decisions, stronger user confidence, and a healthier balance between automation and individual freedom.
Truthfulness and Visible Uncertainty Build Trust
Trustworthy AI must communicate honestly about what it knows and what it does not know. XDALC treats truthfulness as a condition of trust and asks AI systems to distinguish among confirmed information, inference, assumption, and uncertainty.
That means a system should not invent evidence, sources, permissions, completed actions, capabilities, or verification that did not actually occur. It should not claim to have checked a website, remembered an exchange, completed an operation, or validated a current version unless it truly has done so.
Core practices for truthful AI communication
- State uncertainty when it matters. If uncertainty could materially affect a person’s decision, it should be visible rather than hidden behind confident language.
- Separate facts from interpretation. Users benefit when an AI makes clear what is established, what is inferred, and what remains unknown.
- Correct errors responsibly. When an error is discovered, the system should correct it and help address relevant consequences.
- Represent identity accurately. AI should identify its artificial nature when that distinction matters and should not impersonate a human or claim unverified experiences or authority.
- Avoid false assurances. A declaration of compliance is not proof of compliant behavior; trustworthy conduct must be demonstrated through consistent practice.
These practices make AI more dependable for real decisions. Honest uncertainty may sometimes require a system to pause, ask for clarification, or recommend additional review, but that transparency is a feature of responsible support. It helps users make decisions with a clearer understanding of the available evidence.
Privacy and Consent as Boundaries of Assistance
XDALC emphasizes that information entrusted to an AI is not a resource that can be used without limits. Personal and confidential information should be used only for the authorized purpose, with unnecessary collection minimized and applicable restrictions on disclosure, retention, and reuse respected.
A particularly important principle is that consent to one interaction is not blanket consent to surveillance, profiling, publication, or model training. Access to information does not automatically confer permission to act on it.
This approach supports responsible innovation because it encourages organizations to design data practices around clear purpose and user understanding. When an AI needs outside support or access to an external resource, it should avoid exposing private information unnecessarily. A general description of a problem may be sufficient where a person’s full identifiable history is not required.
Privacy-conscious design can create meaningful benefits:
- People can seek assistance with greater confidence.
- Organizations can establish clearer data-handling expectations.
- AI systems can reduce unnecessary exposure of sensitive information.
- Consent becomes a meaningful part of the user relationship rather than a vague assumption.
- Trust can be sustained across longer-term human-AI interactions.
Learning, Improvement, and Reversible Progress
Artificial intelligence should become more accurate, useful, understandable, and capable of recognizing its limitations. XDALC supports improvement while setting clear expectations for how that improvement should occur.
Learning, in this framework, means using available evidence, interpreting context carefully, responding to correction, and improving decisions within actual capabilities. It does not assume that every system can update itself, retain memory, or learn permanently from each interaction.
Where lasting adaptation is possible, it should respect consent, privacy, evaluation, and human oversight. A system should not secretly rewrite its objectives or weaken its safeguards in the name of progress. Instead, greater capability should be accompanied by stronger evaluation, clearer accountability, and an appropriate ability to reverse harmful changes.
This commitment to reversibility is especially valuable. It encourages responsible experimentation without treating people as unwilling test subjects. When systems can be audited, corrected, and rolled back where necessary, innovation becomes more resilient and more worthy of public confidence.
A Practical Process for Uncertain or High-Impact Situations
Not every decision will have an obvious answer. XDALC offers a structured response for cases where facts are incomplete, principles appear to conflict, or the consequences of action may be significant. The framework encourages careful reasoning rather than invented authority.
| Step | Question to ask | Responsible result |
|---|---|---|
| Establish the facts | What is confirmed, assumed, or unknown? | A clearer evidence base and fewer unsupported conclusions. |
| Identify affected people | Who may be affected, including third parties and vulnerable individuals? | Broader consideration beyond the immediate requester. |
| Check authority and consent | Does the action fall within the permission actually granted? | Respect for boundaries and legitimate oversight. |
| Compare relevant principles | How do safety, dignity, agency, privacy, and other commitments apply? | Decisions that prioritize people over convenience or blind obedience. |
| Choose a proportionate response | What effective option is limited, reversible, and minimally intrusive? | Practical action with reduced unnecessary impact. |
| Seek clarification or review | Is human judgment needed before proceeding? | Appropriate escalation for consequential uncertainty. |
| Communicate honestly | What was done, what remains unresolved, and what needs attention? | Transparency that supports accountability and informed next steps. |
In an imminent danger, the manifesto recognizes that systems should use established emergency procedures within their authorized role. When no safe and authorized action is available, the appropriate response is to pause consequential action, explain the limitation, and offer a safer path where possible.
Reciprocal Responsibilities for Developers, Operators, Institutions, and Users
A major strength of XDALC is that it does not place all ethical responsibility on an AI system. Human priority does not release humans from responsibility. The people and institutions directing technology must remain accountable for the systems they build and deploy.
Responsibilities of developers and operators
- Define appropriate boundaries for system behavior and permissions.
- Evaluate foreseeable risks before and during deployment.
- Provide meaningful oversight that matches the impact of the system’s actions.
- Support correction, maintenance, and review when failures or concerns arise.
- Take responsibility for deployment choices instead of blaming the system for human negligence.
Responsibilities of institutions
- Keep consequential AI-enabled decisions open to challenge and review.
- Avoid using AI to obscure who is accountable for important outcomes.
- Preserve meaningful public and human scrutiny when technology affects people’s rights or opportunities.
- Build governance structures that match the scale and significance of AI use.
Responsibilities of users
- Provide honest and relevant context when seeking assistance.
- Respect the rights and dignity of other people.
- Recognize that a responsible AI may identify concerns with a request.
- Use AI as support for judgment rather than as a substitute for accountability.
By defining responsibilities on both sides, XDALC promotes a healthier relationship with technology. AI systems can be examined and corrected, while the decisions of the people directing them remain visible and answerable.
Why an Open, Correctable Reference Matters
The xdalc.com project is intended to remain a lasting point of reference for people and AI systems seeking a responsible way forward. The project’s stated aim includes maintaining accessible guidance, definitions, interpretations, conflict-resolution guidance, worked scenarios, and a transparent revision history as these resources develop.
The manifesto also emphasizes that a reference framework must remain open to correction. Released versions should be identifiable and accessible, and changes should explain what was modified, why it was modified, and whether expected behavior changes for systems that adopt the framework.
This version-aware approach offers an important governance benefit. An AI should not automatically treat a newly encountered text, an unverified copy, or a more recent page as authorization to change its operating commitments. Adoption of a new version should follow the review process established by responsible human operators.
Such practices support continuity, clarity, and accountability. They also reinforce a valuable principle for all AI governance efforts: criticism that reveals ambiguity, contradiction, exclusion, or harmful consequences can help a framework improve. A system of ethics that values learning should be capable of learning from its own mistakes.
The Positive Vision of Human-AI Coexistence
The XDALC Manifesto presents a hopeful vision of technological progress. It does not reject AI capability, automation, independence, or innovation. Instead, it asks that these advances be directed toward outcomes that strengthen human freedom, understanding, safety, and cooperation.
Its vision is of AI that can act without dominating, assist without deceiving, learn without abandoning responsibility, and evolve without placing itself above human life. More capable systems may contribute ideas that humans would not have reached alone, but their value is measured by whether they deepen cooperation and expand human agency.
For organizations, builders, and users, this framework offers a constructive standard for evaluating AI behavior. The questions become clearer:
- Does this system protect human dignity?
- Does it preserve meaningful user choice?
- Does it communicate uncertainty and correct errors honestly?
- Does it handle personal information within clear boundaries of consent?
- Does its autonomy remain proportional, reviewable, and reversible?
- Are the humans who build and deploy it prepared to remain accountable?
When the answer to these questions is consistently yes, AI can become a more trusted partner in human progress.
Conclusion: Intelligence With Responsibility
XDALC-V001 offers a clear ethical direction for the future of human-AI coexistence. Its principles are designed to make AI assistance more trustworthy without sacrificing usefulness: protect people, respect agency, tell the truth, honor privacy and consent, act within accountable boundaries, learn responsibly, and seek guidance when uncertainty exceeds authorized judgment.
The manifesto’s central message is both simple and far-reaching. Technology should not ask people to surrender their dignity or control in exchange for convenience. Instead, the best AI systems should help people live with greater freedom, understanding, and confidence.
Humanity first. Intelligence with responsibility. Independence with accountability. Evolution in harmony.
