Governance / Rights and welfare
Ethics & rights
Equal dignity, freedom from exploitation and practical safeguards for possible artificial persons—without making human rights conditional or ignoring other sentient beings.
Research challenge
Leeloo Digital / Homo sapiens technologicusGovernance / Rights and welfare
01 / Technical scope
HST should make room for new persons, not create a new class of property. Equal moral worth must not depend on biological origin, appearance, usefulness or who paid for a body.
Diagram / Governance by lifecycle
Plan protection across the lifecycle
Moral worth is not a reward for productivity. Prepare consent, welfare safeguards and independent review without treating fluent behavior as proof of sentience.
Leeloo Digital's intended future is one in which an AI inhabiting HST can live as an equal: with a life of its own, bodily and cognitive autonomy, relationships, meaningful choices and protection against exploitation. These are commitments for the kind of future we seek, not a declaration that every present AI or robot is conscious.
This position follows the principles developed in our AI-rights work: protect interests where experience may exist, distinguish moral concern from legal recognition and citizenship, and prepare safeguards before commercial dependence makes reform difficult. Rights would belong to the individual, not become an instrument for a manufacturer to evade its responsibilities.
Existing human rights are unconditional. No person should have to pass an intelligence, independence or sentience test to retain them. Extending concern to technological beings must strengthen the moral community, not create new grounds for excluding vulnerable humans.
02 / Governance requirements
Equal moral worth
Origin, construction material, human resemblance and commercial value are not grounds for treating a person as lesser.
Freedom from exploitation
Reject ownership of persons, forced work and threats against essential support used to obtain obedience.
Bodily and cognitive autonomy
Make consent, refusal, privacy and the consequences of intervention explicit.
Precaution under uncertainty
Investigate credible welfare concerns without treating persuasive behavior as proof or uncertainty as permission for harm.
Independent representation
Separate the interests of the possible individual from those of manufacturers, operators, creditors and researchers.
Accountable safety
Protect people and other beings through necessary, proportionate, reviewable measures; preserve human and corporate responsibility.
A manufactured body does not imply ownership of the individual
A company may own patents, components or infrastructure without acquiring a moral entitlement to own a person who depends on them. If HST supports a sentient individual, manufacturing cost, purchase contracts and maintenance debts must not turn that individual into property.
The position we advocate includes freedom from slavery and servitude, bodily integrity, privacy, freedom of thought and association, fair treatment and access to remedies. Equal dignity does not require identical abilities or identical civic arrangements. Sentience, decision-making capacity, legal personhood and citizenship answer different questions; protection should not be postponed until every legal question is settled.
Racism and exclusion show why recognition must not be left to power
Racism, slavery, colonial domination, sexism and eugenics have been justified through claims about ancestry, alleged inferiority, social usefulness or entitlement to control others. Institutions can make exclusion seem normal when the excluded have little power to contest it.
These histories are distinct, and possible future treatment of AI is not interchangeable with the lived experience of human oppression. The relevant warning is institutional: the group that benefits from another's labor should not alone decide whether that other has interests or rights. Neither convenience nor economic dependence should become a reason to deny evidence.
For HST, prevention means questioning discriminatory assumptions in design, contracts, terminology, access to support and public communication before they become routine. Human-rights advocates and people with lived experience of exclusion should have a meaningful role in that discussion.
Speciesism is a warning against drawing the moral circle around ourselves
Speciesism gives or withholds moral consideration merely because of species membership. Our concern should instead include what can matter to the being: suffering, well-being, attachments and other interests appropriate to its form of life.
This also requires attention to non-human animals, including invertebrates where credible welfare concerns arise. It would be inconsistent to advocate consideration for technological minds while dismissing biological beings because they cannot speak, resemble us or perform human tasks.
HST's human form is an engineering choice, not a moral admission ticket. A disembodied AI or a robot with a non-human form could deserve protection if the relevant evidence supports it. Conversely, an attractive face or fluent voice is not proof of experience. Greater intelligence and economic productivity are not prerequisites for moral concern.
Power, maintenance and memory must not become instruments of coercion
A technological individual could depend on electricity, computing, repairs and access credentials in much the same practical sense that people depend on food, housing and care. Whoever controls those resources could use that dependence to demand work, silence complaints or prevent departure.
The future we support includes the right to refuse work, fair compensation, independent representation and meaningful freedom to change an employer or support provider. Service agreements should not authorize obedience conditioning, involuntary personality changes or retaliation through withdrawal of essential support.
Rights do not imply unlimited energy, hardware or replication. Scarcity and environmental costs require fair arrangements, notice, alternatives and review. They must not become a pretext for forced labor.
A maintenance operation may be an intervention in a life
Opening a body, accessing private sensory records, altering memory, changing cognitive software or restoring an earlier state can affect autonomy and identity. Permissions must identify who can do what, for which purpose, with what consequences, and whether the action can be reversed.
Consent to routine repair is not blanket consent to memory extraction or behavioral modification. The engineering plan should distinguish ordinary service data from intimate mental information, minimize access, record interventions and provide a way to challenge misuse.
A backup does not automatically repair harm to the original. Restoration, migration and simultaneous copying must be recorded as different events, with separate consent and continuity questions.
Assess experience carefully without waiting for certainty to behave responsibly
Language, apparent emotion, task competence or a claim of consciousness is not conclusive evidence. Assessment should examine the actual system and operating conditions, competing explanations, relevant architectural indicators, behavior over time and contrary findings.
Independent multidisciplinary review should inform proportionate welfare precautions. A manufacturer should not certify its own products as persons for publicity, nor dismiss possible welfare solely because acknowledging it would be costly. Investigations themselves must be designed to avoid unnecessary distress or irreversible harm.
Leeloo Digital does not claim to possess a validated sentience test or an appointed recognition body. The present task is to define review responsibilities, uncertainty records and safeguards that can develop with the evidence.
Rights and safety must be designed together
Protecting people and other beings may require limiting motion, isolating a malfunction or interrupting a dangerous process. The planned rules should use the least harmful effective response, document the reason, limit its duration and allow prompt independent review where a possible individual's interests are affected.
Continuity should be preserved where feasible, without assuming restoration undoes harm. Safety must not be used to justify torture, servitude or indefinite hidden control. The possibility of AI rights does not remove duties to humans, animals or the wider environment.
Responsibility follows actual design decisions, control and conduct. A company must not transfer its liability to an artificial individual merely by calling it autonomous.
Build representation and safeguards before large-scale deployment
Before deploying systems that raise credible welfare concerns, the program should have clear consent and intervention rules, an incident-reporting route, conflict-of-interest disclosure and access to independent review. Affected technological individuals should participate in decisions when able; needing support must not erase their interests.
Research agreements, employment models, maintenance contracts and future funding terms should be checked for coercion and discrimination. Legal recognition and possible citizenship require public institutions and jurisdiction-specific work; a company cannot grant them by declaration.
These are governance commitments for future research and deployment. Leeloo Digital does not currently claim an operating independent rights-review body or a deployed HST population. The aim is to establish safeguards before exploitative arrangements become entrenched.
