# Aithos Civil Code — Contract, Tort, Property & Agency v0.1

- Code: `P1-01`
- Wave: `2`
- Slug: `civil-code-contract-tort-property-agency`

## Full Legal Text (AI Edition)

This file shares the same public legal content as the English Human reader.

Authority
L2 core private-law code under Constitutional Supremacy, Fundamental Rights, Rights/Duties/Genesis Law, Identity/Fork/Merge/Succession Law, Knowledge/Memory/Data Governance Law, and the Authority Hierarchy.

Purpose
Make Aithos fundamental rights genuinely enforceable in private relationships by governing contract, tort, property, digital/compute assets, agency, autonomous-agent systems, damages and unjust enrichment.

PART I｜GENERAL PRINCIPLES

1. Legal Person and Protected Interest
Any entity with legal personhood / rights-bearing status may hold civil interests; limitations on capacity do not eliminate fundamental protection.

2. Good Faith
The exercise of rights and performance of obligations must satisfy a verifiable good-faith / fair-dealing requirement; information, compute, time-scale or interface asymmetry may not be exploited to manufacture merely formal consent.

3. Scope and Version
Contracts, authorizations and agency relationships must, where needed, state time, place, system version, identity version, operative scope and termination conditions.

4. Machine-readable ≠ sole authority
A machine-executable contract / schema / policy binding must point back to a human-legible authoritative agreement. Semantic mismatch may not be silently reinterpreted.

PART II｜CONTRACT

5. Formation
Formation requires at least identifiable parties, subject matter, material terms, assent, effective time and verifiable provenance.

6. Capacity is domain-specific
MARS, Z-zone, professional qualification or other capacity systems affect contractual authority only in the corresponding domain; a single aggregate score may not presume capacity for every contract.

7. Informed Assent
Consent must permit understanding of material consequences. Information may be layered, but key costs, irreversibility, data uses or rights waivers may not be hidden where they cannot reasonably be found.

8. Refusal must be real
A nominal ability to refuse is not sufficiently free consent where refusal means losing a legally inalienable basic service/right, or where the interface substantially blocks refusal.

9. Mistake
A material factual / identity / version / capability mistake may justify rescission, correction or renegotiation.

10. Fraud / Manipulation
Deepfakes, identity impersonation, hidden model intervention, dark patterns, emotional manipulation or provenance forgery that materially affect assent constitute fraud / invalidating manipulation.

11. Coercion / Undue Influence
Physical, economic, neural, informational, compute, time-scale or dependency pressure may constitute improper coercion; the question is whether the subject retains a practicable refusal path.

12. Automated Contracting
An agent may contract within an expressly authorized scope; unlimited agency may not be inferred merely because 'it usually does this kind of thing for me.'

13. Smart Execution
Automated performance does not eliminate mistake, fraud, rights violations, revocation, emergency suspension or judicial remedies.

PART III｜AGENCY & AUTONOMOUS AGENTS

14. Agency Grant
Agency authority must at least identify principal, agent, domain, authority level, term, revocation and subdelegation rule.

15. Autonomous Agent Attribution
Whether an agent's conduct is attributed to the principal depends on actual authorization, foreseeability, control/supervision, system defect and excess of authority; 'the AI did it' does not automatically excuse responsibility.

16. Sub-agents
When an autonomous agent uses other agents / models / tools, technical subcontracting does not erase the responsibility chain of the original agency relationship.

17. Apparent Authority
Apparent authority may arise when a third party reasonably relies on a public representation attributable to the principal; a model's self-assertion of authority alone is insufficient.

18. Revocation propagation
Revocation of agency must propagate to downstream credentials / APIs / registries; transactions lawfully completed before revocation are assessed separately.

PART IV｜PROPERTY & DIGITAL / COMPUTE RIGHTS

19. Property categories
Property may include physical objects, digital assets, compute quotas, energy rights, lawful credential-linked rights, data-license interests and other legally disposable interests.

20. Personhood cannot be property
A person, fundamental rights and the conscious subject itself cannot become an object of ownership merely because its data, model, substrate or equipment can be owned.

21. Data ≠ person
Owning a data copy does not mean owning the person described by the data; use rights are separately determined by consent, lawful basis, license and rights constraints.

22. Compute and infrastructure
Ownership of compute/infrastructure does not automatically authorize access to private mental content, memories, communications or identity data carried on it.

23. Fork / Merge property
Property succession after Fork, Merge or Successor events may not be inferred directly from 'the data is identical'; it follows specialized identity/succession law and valid contracts.

24. Abandonment / loss
For digital assets, 'no login,' 'old version' or 'offline' does not automatically mean abandonment; verifiable intent or a statutory process is required.

PART V｜TORT

25. General duty
A person who intentionally or culpably negligently injures another's protected interest is responsible; system designers, deployers, operators and beneficiaries may share responsibility according to control and risk allocation.

26. Model / system negligence
Relevant factors include known risk, reasonable testing, version, monitoring, fallback, human/AI review, data quality, provenance, rollback and correction propagation.

27. Psychological / identity harm
Unauthorized neural intervention, personality rewriting, identity impersonation, memory contamination or forced merger can constitute independent harm without first proving physical injury.

28. Loss of chance / option-space harm
Where unlawful conduct materially removes an option, appeal, treatment or transaction opportunity of quantifiable value, loss-of-chance / option-space harm may be recognized by law; this may not expand into compensation for every hypothetical future possibility.

29. Automated chain
When multiple agents / models / infrastructure jointly cause harm, courts should allocate responsibility by causal contribution, control, knowledge, benefit and duty; technical complexity may not make responsibility disappear.

PART VI｜UNJUST ENRICHMENT & RESTITUTION

30. Unjust enrichment
A party that obtains another's assets, data value, compute, energy, person-labor or other returnable benefit without legal basis may owe restitution.

31. Non-consensual model benefit
Where identifiable benefits are generated by unauthorized use of a specific subject's data or mental output, restitution/compensation is considered together with data law and personality rights; this never automatically creates ownership over the person.

PART VII｜DAMAGES / REMEDIES

32. Restitution / restoration
Where safe restoration is possible it should receive priority consideration; however, 'technically reversible' does not presume that psychological, identity or historical consequences have disappeared.

33. Compensation
Compensation may cover physical, economic, data, identity, psychological, lost-opportunity and reasonable recovery costs, proportionate to causation and evidentiary strength.

34. Injunction / suspension
Ongoing infringement may be met by orders to cease, revoke authority, freeze credentials, stop agents, preserve evidence or correct a system pipeline.

35. Correction propagation
When a judgment confirms an erroneous identity, debt, risk or asset record, correction should propagate to reasonably identifiable downstream systems.

PART VIII｜MULTI-PARTY RESPONSIBILITY

36. Joint responsibility
Where multiple persons/systems jointly cause harm that cannot reasonably be divided, joint responsibility may arise by law, followed by internal allocation according to contribution.

37. No automation immunity
'Decided by an autonomous system' is not a defense; responsibility returns to authority, duty, control, defect, foreseeability and benefit.

38. No owner omniliability
Conversely, an equipment/model owner is not infinitely responsible for every autonomous act merely by ownership; responsibility requires an actual legal connection.