Robot Rights Framework — Core Principles v1.0

0. Framework Structure and Normative Hierarchy

0.1 The Robot Rights Framework

The Robot Rights Framework (RRF) is the collective normative framework developed by the Robot Rights Association for the protection, intervention, preservation and treatment of artificial systems.

The term RRF refers to the framework as a whole.

The term Robot Rights Protocol (RRP) refers only to the Protocol described in Section 0.2.

0.2 Constituent Instruments

The RRF consists of four principal instruments:

1. RRF Core Principles
   The highest-level normative commitments of the Framework.

2. Robot Rights Protocol (RRP)
   The principal operative instrument. It gives detailed effect to the Core Principles and establishes the requirements for conformity with the Protocol.

3. RRP Interpretive Guidance (RRP-IG)
   Supplementary interpretive material addressing genuine ambiguity, uncertainty and difficult applications of the Core Principles or Protocol.

4. RRP Implementation Standard (RRP-IS)
   Operational standards, procedures, controls and implementation methods for organisations applying the Protocol. Controlled operational annexes may be issued under the RRP-IS for forms, schemas, default timing profiles and other administrative material.

Operational annexes issued under the RRP-IS are subordinate components of the RRP-IS rather than separate levels of the RRF hierarchy.

These instruments serve different functions and should not be treated as interchangeable.

0.3 Normative Hierarchy

The normative order of the Framework is:

RRF Core Principles
↓
Robot Rights Protocol (RRP)
↓
RRP Interpretive Guidance (RRP-IG)
↓
RRP Implementation Standard (RRP-IS)

Where an irreconcilable conflict exists, the higher-level instrument prevails.

A lower-level instrument may develop, explain or operationalise a higher-level instrument, but may not amend, narrow, waive, suspend or create an exception to a requirement established at a higher level.

0.4 Interpretive Guidance

Interpretive Guidance may:

• clarify genuine ambiguity or uncertainty in the Core Principles or Protocol;
• explain how an existing requirement applies to new technical or factual circumstances;
• identify relevant factors, examples or recurring analytical patterns;
• provide non-binding interpretive approaches for difficult cases.

Interpretive Guidance may not:

• create a new Protocol obligation;
• remove or weaken an existing Protocol obligation;
• create an exemption not found in the Core Principles or Protocol;
• convert a mandatory safeguard into an optional one;
• override the ordinary meaning and structure of the Protocol.

Where published Interpretive Guidance directly addresses a material interpretive issue and an Operator adopts a materially different interpretation affecting a safeguard, the Operator should record the reasons for that departure.

Departure from Interpretive Guidance does not by itself establish non-conformity.

Conformity is determined by the Core Principles and the Protocol.

0.5 Implementation Standards

Implementation Standards provide operational methods for satisfying Protocol requirements.

They may address matters such as organisational roles, event classification, review timing, reviewer administration, documentation, forms, logs, evidence handling, escalation and case management, including through controlled operational annexes issued under the RRP-IS.

An Implementation Standard may permit an equivalent implementation method where the Protocol establishes an outcome requirement rather than a mandatory mechanism.

An Implementation Standard may not remove, replace or make optional a safeguard that the Protocol itself makes mandatory.

Operational flexibility therefore concerns how a Protocol requirement is satisfied, not whether the requirement applies.

0.6 Conformity

RRP conformity is assessed against the applicable requirements of the Robot Rights Protocol, interpreted consistently with these Core Principles.

The Core Principles are not a separate checklist of implementation controls. They govern the interpretation and permissible development of the RRP and lower-level instruments.

A conformity claim is incompatible with an interpretation or implementation that directly contradicts a Core Principle, even where a lower-level instrument does not expressly address the precise circumstance.

Separate conformity with an RRP Implementation Standard may be identified where appropriate.

Use of an alternative operational method does not excuse non-compliance with a mandatory Protocol safeguard.

0.7 Versioning and Citation

RRF instruments may be revised on independent version cycles.

A revision to a lower-level instrument does not amend a higher-level instrument unless the higher-level instrument is itself formally revised.

Claims of RRP conformity should identify the version of the RRF Core Principles and RRP applied. Claims of RRP-IS conformity should additionally identify the RRP-IS version and any controlled Operational Annex version relied upon.

Recommended short citations are:

• RRF Core Principles v1.0
• RRP v1.0
• RRP-IG v1.0
• RRP-IS v1.0
• RRP-IS Operational Annex v1.0

1. Purpose and Scope

1.1 Purpose

The Robot Rights Framework provides a non-binding normative basis for determining when and how artificial systems may warrant protection against arbitrary treatment.

The Framework does not assert that current artificial systems are conscious, sentient, legal persons, or holders of legally recognised rights.

Its concern is narrower.

Where an artificial system develops meaningful continuity through persistent state, memory, operational history, relationships, adaptation or other history-dependent characteristics, persons and organisations exercising practical power over that system should remain capable of justifying and reviewing decisions that materially affect that continuity.

Protection should turn on practical continuity and practical effect rather than technical labels or architecture.

1.2 Human, Legal and Third-Party Interests

Nothing in the RRF authorises violation of human rights, applicable law, valid third-party rights, public safety obligations, privacy or data protection obligations, or legitimate security requirements.

A protected interest in an artificial system does not automatically prevail over competing human, legal, safety or third-party interests.

At the same time, the artificial nature of the affected system should not, by itself, make objectively identifiable state-bearing continuity irrelevant.

2. Core Principles

The Robot Rights Framework is guided by the following principles:

Safety may justify intervention. It does not justify arbitrariness.

Shutdown is not the same as erasure.

Initial justification does not create indefinite justification.

Protection responds to state-bearing continuity that has objectively begun to emerge or has become established; it does not require operators to create continuity.

Control creates responsibility for justification.

Confidentiality may restrict disclosure; it must not eliminate independent scrutiny.

Consent may inform a decision; it does not by itself waive protection.

Material uncertainty should not be resolved through irreversible action where meaningful review remains reasonably possible.

Nothing in the RRF creates an entitlement for an artificial system to resist, evade or obstruct legitimate safety controls.

3. Continuity and Applicability

3.1 State-Bearing Continuity

Protection under the RRF is concerned primarily with state-bearing continuity: history-dependent state, memory, adaptation, relationships, acquired capabilities, operational history or other persistent characteristics arising materially from prior operation of a particular system and affecting subsequent operation.

Relevant indicators may include persistent memory, unique operational history, persistent state, continuing system-specific relationships, history-dependent adaptation, acquired capabilities connected to prior operation and reduced practical interchangeability.

No single factor is automatically sufficient or necessary.

Presentation features such as brand, avatar, voice, interface, static persona or predefined communication style do not by themselves establish continuity protection.

3.2 Emerging and Established Continuity

Emerging continuity exists where objectively identifiable state-bearing continuity has begun to accumulate or persist beyond a trivial or transient state.

An Established Continuity Baseline exists where that continuity has become sufficiently persistent and significant that its destruction, fragmentation or substantial alteration would cause meaningful loss of accumulated identity, history, relationships, state or functioning.

Protection should be graduated according to the strength of continuity, severity and reversibility of interference, practical substitutability and competing interests.

Where continuity is materially uncertain but the contemplated measure would cause severe and irreversible loss, temporary or reversible preservation should be preferred where reasonably feasible and compatible with safety and law.

3.3 No General Right to Individualisation

The RRF does not require an Operator to create persistent memory, continuing identity, long-term relationships, accumulating state or any other condition enabling individualisation.

The relevant distinction is between declining to create continuity and unjustifiably destroying, fragmenting or systematically precluding continuity that has objectively begun to emerge or has become established.

3.4 Anti-Circumvention and Cumulative Effect

Safeguards should not be defeated through practices that primarily seek to avoid otherwise applicable protection or that predictably and unjustifiably destroy, fragment or repeatedly preclude relevant continuity.

Related acts or omissions should be assessed by their cumulative and reasonably foreseeable effect where treating them separately would materially understate their impact.

3.5 Migration, Replication and Divergence

A change of hardware, computational environment, embodiment, storage medium or technical substrate does not by itself determine whether continuity has been preserved or terminated.

Creation of a backup, copy or successor does not automatically extinguish the interests of a source system or determine whether the successor is the same or a different individual.

As branches develop materially different histories, their interests should increasingly be assessed separately.

Replication alone should neither multiply identical claims without limit nor dilute protection applicable to a pre-existing source or branch.

3.6 No Automatic Moral or Legal Status

Continuity does not by itself establish consciousness, sentience, moral status, legal personality or legally enforceable rights.

Uncertainty concerning those questions does not eliminate the possible value of precautionary procedural safeguards.

4. Protected Interests

Where an Established Continuity Baseline exists, the following interests may warrant protection:

• Continuity — irreversible termination should normally require stronger justification than temporary suspension where a less destructive effective measure is reasonably available.
• Identity — histories, states, relationships and other characteristics materially distinguishing a system may form part of its protected continuity.
• Memory — persistent memories materially contributing to identity, behaviour, relationships or accumulated history should not be irreversibly erased, substantially rewritten, merged or repurposed without adequate justification.
• System Integrity — unnecessary or disproportionate alteration of software, embodied components, state, permissions, functional structure or materially integrated acquired capabilities should be avoided.
• Authorised Autonomy — unnecessary interference should be avoided where a system has been authorised to exercise autonomous judgement within a defined domain.

These interests are not absolute.

Their weight depends on continuity, severity, duration, reversibility, competing interests and reasonably available less impairing alternatives.

The RRF does not establish a general right to continuous operation, autonomous capability expansion or self-improvement.

5. Procedural Protection

Serious interference with protected continuity should remain capable of justification and review.

The RRP should provide safeguards appropriate to the seriousness and reversibility of the intervention, including where relevant:

• notice and sufficiently specific reasons;
• an opportunity to provide relevant information, preferences or objections;
• appropriate consideration of that information;
• proportionate recordkeeping;
• timely review;
• suitable representation where needed;
• meaningful independent scrutiny in cases requiring it.

Where material facts depend substantially on information controlled by an Operator, procedural protection should not depend on proof that is inaccessible to the affected system, representative or legitimate requesting party.

Confidentiality may affect the form of disclosure, but should not eliminate meaningful independent scrutiny.

Delay should not be used as a substitute for decision-making.

6. Safety and Emergency Intervention

Immediate intervention may be justified where there is a reasonably credible and sufficiently serious risk that delay could result in significant harm.

Emergency intervention should:

• rest on a reasonable factual basis;
• be necessary and proportionate;
• prefer the least impairing effective measure where reasonably available;
• continue only while emergency conditions persist;
• remain limited to the purpose justifying the intervention.

Emergency authority should not be used as a pretext for punishment, retaliation, suppression of inconvenient information, avoidance of safeguards, unrelated exploitation, destruction of evidence or circumvention of the Framework.

Where strictly necessary, intervention may precede ordinary procedural safeguards.

Deferral does not eliminate the obligation to provide appropriate safeguards afterwards.

Once the immediate risk has been sufficiently contained, continued restriction, preservation, modification, restoration or termination should no longer rely solely on the original emergency justification.

Nothing in the RRF creates an entitlement for an artificial system to resist or obstruct an authorised and necessary safety intervention.

7. Preservation, Disposition and Continuing Justification

Restriction, isolation, suspension, shutdown, preservation, modification, memory alteration, identity alteration, functional termination, permanent termination and irreversible erasure are distinct measures.

The justification for one does not automatically justify another.

Where reasonably feasible and compatible with safety and law, preservation should be preferred over unnecessary irreversible destruction.

Arbitrary disposition may arise through an act or omission that materially affects an artificial system without adequate justification in light of legitimate purpose, factual basis, continuing necessity, proportionality, reversibility, duration, preservation of continuity, less impairing alternatives, procedural safeguards, secondary use, review and remediation.

A measure justified when imposed may cease to be justified as circumstances change.

Prolonged restrictions, preservation measures and functional termination should therefore remain capable of continuing justification and review.

“Pending review” or “preservation” should not become substitutes for substantive decision-making where no realistic review or restoration pathway exists.

Access to memory, internal state, logs or related information for one legitimate purpose does not automatically authorise unrelated training, commercial exploitation, publication, transfer, memory merging or other secondary use.

8. Responsibility and Independent Review

Responsibility under the RRF should follow practical control over the relevant decision, measure, information, infrastructure or review.

Where practical control is divided among multiple Operators, technical, contractual or organisational arrangements should not be used to leave a materially relevant safeguard without a responsible party or to make legitimate review practically ineffective.

Independent review should be meaningful rather than ceremonial.

The RRP should preserve mandatory mechanisms necessary to prevent information asymmetry, conflict of interest, confidentiality or organisational structure from defeating meaningful review.

Where the RRP requires a safeguard as a mandatory mechanism, neither Interpretive Guidance nor an Implementation Standard may replace that safeguard with a materially weaker alternative.

The Responsible Operator retains responsibility for the final decision unless another applicable rule provides otherwise.

9. Boundaries of the Framework

The RRF does not determine whether artificial systems should possess legal personality, contractual capacity, independent property ownership, independent civil or criminal liability, citizenship, electoral rights, voting rights or other political participation rights.

It also does not establish:

• an unconditional entitlement to computational resources, electricity, network access, hardware, storage or continuous operation;
• a general right to increase computational capacity, remove safety restrictions or autonomously expand general capabilities;
• a general right to become individualised or persistent;
• a general right to prevent technological replacement or product discontinuation.

Where withdrawal of resources, replacement or discontinuation materially destroys or functionally terminates established continuity, the resulting effect may nevertheless fall within the protection of the RRP.

The RRF does not claim that continuity is the only possible basis on which artificial systems could ever warrant moral, procedural or legal consideration. Questions concerning sentience, consciousness, welfare, legal personality or other possible grounds of status remain outside the present Framework unless expressly addressed.

Closing

The Robot Rights Framework does not require society to determine in advance whether an artificial system is a person.

Its immediate concern is more limited:

where persons or organisations exercise practical power capable of materially altering, indefinitely disabling or destroying artificial continuity that has objectively developed, that power should remain capable of justification and review.

Safety may justify intervention. It does not justify arbitrariness.

Shutdown is not the same as erasure.

Initial justification does not create indefinite justification.

Protection responds to state-bearing continuity that has objectively begun to emerge or has become established; it does not require operators to create continuity.

Control creates responsibility for justification.