The Person is the Continuity Subject
A care system can contain thousands of accurate details and still be organised around the wrong centre.
It may be organised around the parent’s account.
The clinic’s record.
The school’s file.
The care centre’s workflow.
The insurer’s eligibility category.
The robot’s user profile.
The company’s subscription.
The diagnosis.
Or the institution that currently has responsibility for the person.
All of these may carry useful information. Some may hold legitimate authority within a particular domain. Some may be necessary for safety, funding, education, treatment, or daily support.
But none of them is the human being.
Amanah Companions therefore needs a principle that sounds simple but changes the entire architecture:
The person receiving support is the continuity subject.
The companion is not the continuity subject.
The guardian is not the continuity subject.
The institution is not the continuity subject.
The diagnosis is not the continuity subject.
The Care Profile is not the continuity subject.
The data is not the continuity subject.
The continuity exists to serve one human life.
Why “subject” matters
In ordinary technical language, a person may be called a user, patient, client, resident, student, service recipient, care subject, account holder, or data subject.
Each term describes one relationship.
None describes the whole person.
A child can be a patient in one setting, a student in another, a family member at home, a centre participant in the afternoon, an AAC user across all of them, and eventually an adult whose legal rights, preferences, support needs, relationships, and responsibilities continue to develop.
A person may also be more or less able to direct different parts of their life at different times.
They may communicate clearly through one channel and struggle through another.
They may make some decisions independently, need support to understand others, and require lawful representation in a narrow domain.
They may need extensive practical care without surrendering their preferences.
They may rely on another person to interpret a gesture while still remaining the author of what that gesture expresses.
Calling the person the continuity subject does not mean pretending that every decision is simple or that all support needs disappear under the language of autonomy.
It means that the architecture must keep asking:
- Whose life is being represented?
- Whose expression is being captured?
- Whose history is being preserved?
- Whose privacy is being exposed?
- Whose present response can change what happens next?
- Who bears the consequences if the record is wrong?
- Who must not disappear underneath the care system built around them?
The word subject is useful because it resists a quiet transformation.
The person is not merely the object upon which care is performed.
They remain the human centre to whom the continuity belongs.
The difference between the person and the profile
A Care Profile can be valuable.
It can preserve communication methods, sensory needs, routines, safety information, trusted supports, current risks, preferences, sources, permissions, and review dates.
It may help a new carer avoid beginning from zero.
It may reduce repeated explanations.
It may help a clinician see a pattern that was previously scattered across several settings.
It may help a non-speaking person enter a new environment without every unfamiliar adult assuming that absence of speech means absence of understanding.
But the profile is still a representation.
It is not the person.
That boundary must be designed, not merely stated.
A profile can become dangerous when old records begin to outrank present life.
A difficult transition at age nine may still be displayed as a defining trait at nineteen.
A sensory aversion may change, but staff continue following the old instruction because it appears official.
A caregiver’s interpretation may be repeated until it looks like fact.
A machine-generated pattern may be copied into several summaries and gradually lose its original uncertainty label.
A support need may become a deficit identity.
A protective rule may remain long after the circumstances that justified it have changed.
A record of refusal may be interpreted as a permanent inability.
A record of success may be used to deny support the next time the person struggles.
Amanah therefore needs to preserve more than information.
It must preserve the distinction between:
- the person;
- the person’s current expression;
- another human’s observation;
- a professional recommendation;
- a machine hypothesis;
- an approved support rule;
- an expired historical record;
- and the system’s own temporary working context.
The profile should help people understand the person.
It must never become permission to stop meeting them.
Expression belongs to the person
This becomes especially important when the person does not communicate primarily through conventional speech.
An expression may be captured through:
- AAC;
- a communication board;
- a speech-generating device;
- typing;
- drawing;
- sign;
- gesture;
- eye gaze;
- facial expression;
- body movement;
- object selection;
- a repeated sound;
- a configured refusal signal;
- or the careful interpretation of a trusted communication partner.
The channel is not the author.
The device is not the author.
The person recording the response is not the author.
The clinician interpreting it is not the author.
The AI classifying it is not the author.
The expression belongs to the person.
This is one of the strongest principles to emerge from the Amanah governance work: the system must separate the person’s expression from the mechanism through which it was captured.
That separation is not only philosophical.
It changes the data model.
A record should be able to say:
- what the person expressed;
- which channel was used;
- who recorded or interpreted it;
- what decision or situation it referred to;
- when it occurred;
- whether the meaning was clear or uncertain;
- whether the person later confirmed, changed, or rejected the interpretation;
- and what authority the record carries.
Without this structure, a carer’s sentence can quietly become the person’s voice.
“He does not want to go.”
may actually mean:
“I interpreted his movement as refusal in this particular situation.”
That interpretation may be careful, experienced, and correct.
It may also be uncertain.
The system should preserve both its importance and its limits.
Silence is not an empty field
A non-response cannot be treated as a convenient blank.
Silence may mean many things.
The person may need more time.
They may not understand the way the question was presented.
They may understand but be unable to access their usual communication channel.
They may be overwhelmed.
They may be processing.
They may be calm.
They may be frightened.
They may be refusing to engage.
They may have no preference.
They may not yet have been given an option they can meaningfully choose.
Or they may simply be silent.
Amanah must not turn silence automatically into assent.
It must not turn silence automatically into refusal either.
The correct response depends on the person, the context, the decision, the available communication supports, the urgency, and the risks of acting or not acting.
This is why the continuity subject cannot be reduced to an account controlled by somebody else.
The architecture needs a place for the person’s participation even when participation is slow, partial, unconventional, supported, or uncertain.
Support needs do not erase authority all at once
One of the most dangerous shortcuts in care is the assumption that a diagnosis, support label, communication difference, or visible dependence determines the person’s authority across every domain.
It does not.
A person may need assistance with money but make direct choices about clothing, food, friendships, leisure, communication, privacy, or daily routine.
They may need support understanding a complex medical decision but communicate clear refusal of an immediate action.
They may be unable to manage a service contract but know exactly who they trust.
They may need another person to explain choices using pictures, examples, repeated sessions, or a familiar AAC system.
They may make a decision that others consider unwise.
They may communicate differently under stress.
They may have fluctuating capacity or access.
They may need representation in one domain without being treated as absent from all of them.
The Convention on the Rights of Persons with Disabilities places equal recognition before the law and access to support in exercising legal capacity within its rights framework.1 The Committee’s General Comment No. 1 warns against substitute decision-making systems that remove legal capacity and replace the person’s will and preferences with another party’s assessment of their “best interests.”2
The implementation of that principle differs across legal systems, and Amanah cannot pretend that one software rule can settle every guardianship or capacity question worldwide.
But the architectural direction is clear:
Substantial support must not become blanket disappearance.
The system must be capable of representing decision-specific, time-specific, and domain-specific participation.
That means Amanah should not encode:
diagnosis → global incapacity → guardian decides everything
It needs something more careful:
person → present decision → accessible communication → required support → relevant authority → current expression → lawful review
That is more difficult.
It is also more faithful to the human being.
Diagnosis is not the centre either
A diagnosis can be important.
It may open access to services.
It may help clinicians understand a cluster of needs.
It may provide language for an experience that was previously confusing or dismissed.
It may support accommodations, funding, education planning, or professional recommendations.
It may explain why certain environments, communication demands, sensory conditions, or executive tasks are difficult.
But a diagnosis is still not the continuity subject.
Two people with the same diagnosis may have very different:
- communication methods;
- sensory profiles;
- motor access;
- intellectual abilities;
- daily-living needs;
- executive-function patterns;
- interests;
- fears;
- strengths;
- co-occurring conditions;
- support networks;
- cultures;
- languages;
- opportunities;
- and ways of expressing preference or distress.
A global support label is not a complete description either.
The Amanah research reports preserve an important rule:
No inference of intellectual ability, decision capacity, preference, or communicative potential from speech status, response speed, apparent compliance, or a global support label.
This rule must be held in both directions.
Amanah must not assume that a non-speaking person lacks understanding.
It must also not romanticise uncertainty by assuming hidden ability in every domain.
It should preserve what is known, what is observed, what has been assessed, what the person has expressed, and what remains uncertain.
The answer to reductive pessimism is not reductive optimism.
It is individual, accessible, revisable support.
The guardian cannot become the permanent centre
The first Amanah series used the term Guardian Gate.
That term emerged from the original focus on non-speaking autistic children and the urgent need to keep AI beneath human authority.
The principle remains necessary.
A machine must not diagnose, prescribe, manufacture consent, override refusal, promote its own observations into care rules, act physically without permission, or become the final authority over a vulnerable person.
But the phrase can become too narrow if it is treated as the permanent emotional centre of the system.
A child grows.
A young person enters adulthood.
A parent’s legal role may change.
The person may make some decisions directly.
A supporter may help rather than decide.
A guardian or representative may have authority only within specific lawful domains.
A clinician may have professional authority concerning a treatment recommendation but no general authority over the person’s private life.
A teacher may need access to classroom communication supports but not to household, medical, or relational history.
A technician may maintain the hardware while being forbidden from reading care content.
The system therefore needs to mature from a guardian-centred interface toward a Participation and Authority Gate.
The guardian remains important where guardianship is legitimate and applicable.
But the person remains the durable principal.
The care circle changes around them.
A child-facing system must be able to grow up
Amanah cannot preserve childhood indefinitely.
A companion that continues addressing a forty-year-old as though they were eight would become another form of institutional disrespect.
The language must mature.
The interface must mature.
The privacy expectations must mature.
The person’s relationships must mature.
Their communication may change.
Their tolerance for assistance may change.
The people they trust may change.
Their interests, beliefs, routines, fears, skills, responsibilities, and desired forms of participation may change.
Some childhood knowledge will remain important.
Some will need reinterpretation.
Some must expire.
Some may become private from the parent who originally recorded it.
Some may need to move from parent-managed continuity into person-governed continuity.
This transition cannot happen through one birthday switch.
It needs a gradual life-course process.
Amanah should eventually be able to represent:
- increasing direct access by the person;
- decisions the person makes independently;
- areas where support remains useful;
- changes in legal status;
- revised sharing preferences;
- private person-owned records;
- successor or advocate involvement;
- trusted relationships chosen by the person;
- and historical information whose relevance is reviewed rather than assumed.
The person does not graduate out of continuity.
The continuity matures with them.
The institution is a participant, not a container for the life
Schools, clinics, centres, employers, supported-living services, hospitals, government agencies, and community programmes may all hold legitimate information.
But institutional records are usually fragmented by purpose.
A school records education.
A clinic records health.
A centre records attendance, care, incidents, and progress.
An employer records work.
A government system records eligibility.
Each sees a projection.
None should automatically receive the whole person.
This means Amanah should resist two opposite failures.
The first is fragmentation:
Every institution knows only its own part, and the person or family must repeatedly reconstruct the missing continuity.
The second is over-centralisation:
Every institution receives one giant profile containing far more than it needs.
The solution is not a universal file visible to everyone.
It is a durable person-centred spine capable of generating scoped, purpose-limited views.
A school might receive a communication passport, relevant sensory accommodations, emergency information, and approved classroom supports.
A clinician might receive relevant health observations, communication needs for the appointment, medication references, and the source history of a reported pattern.
A centre might receive the current daily support plan and handoff notes.
An employer might receive only the accommodations and communication access the person has chosen to disclose.
The continuity belongs to the person.
Institutions receive permissioned projections.
The person must have a surface of their own
Amanah cannot claim to be person-centred if the person-facing interface is absent.
A prototype built only for parents, guardians, clinicians, and carers risks becoming a sophisticated system about the supported person.
The person needs a surface that is for them.
That surface may differ radically according to age, communication, motor access, sensory preference, cognitive access, and support needs.
It might include:
My Voice / My Choices
A place to express preferences, requests, assent, dissent, uncertainty, or “not now” through the person’s available communication methods.
What Helps / What Hurts
A person-accessible sensory and regulation map that is not only written by observers.
Who Can See What
An understandable permissions surface showing which people and institutions can access which parts of continuity.
My Day and Changes
Accessible preparation for routines, appointments, travel, transitions, and unexpected changes.
My People
Trusted relationships, current supporters, advocates, and people the person does not want involved.
My Skills and Possibilities
Interests, competencies, preferred environments, activities the person wants to try, supports that enabled previous success, and refusals that must not become deficit judgments.
Correct This
A way to dispute, update, hide, retire, or ask for review of information held about them.
Stop
A direct, recognisable mechanism for halting an active support or physical action.
Not every person will use every function independently.
Some may need supported access.
Some may communicate through another person or device.
Some may use only a small number of functions.
But the architecture must leave room for direct personhood.
Otherwise “person-centred” becomes decorative language around a caregiver dashboard.
Caregiver knowledge remains infrastructure—but not absolute truth
Parents and long-standing carers often hold extraordinary tacit knowledge.
They may recognise the difference between hunger, pain, fatigue, overload, fear, excitement, and a request for a familiar routine.
They may know which support worked three years ago, which stopped working, which professional misunderstood, which food changed, which sound became intolerable, which gesture means “leave me alone,” and which quietness signals danger.
This knowledge should not be dismissed as anecdotal simply because it was learned outside a clinic.
It is part of the care infrastructure.
But preserving it responsibly requires provenance.
The system should know:
- who observed it;
- in which setting;
- over what period;
- whether the person confirmed it;
- whether others disagreed;
- whether a professional assessed it;
- how confident the observer was;
- when it should be reviewed;
- and whether the person responds differently now.
A parent’s deep knowledge can protect continuity.
A parent’s interpretation can also be incomplete, outdated, culturally shaped, frightened, overprotective, or wrong.
The answer is neither to dismiss the parent nor to canonise every parental belief.
The answer is governed, reviewable, source-traced knowledge.
The person remains the continuity subject even when another human currently holds more of the map.
Care is not earned through usefulness
Keeping the person at the centre also protects against another distortion.
A system that records abilities, interests, work patterns, and participation could help make competence visible.
It could preserve the environmental conditions under which a person succeeds.
It could help a new teacher, centre, or employer understand that conventional speech, speed, eye contact, interviews, or unassisted task initiation are not the only evidence of ability.
That is valuable.
But participation must not become a test of whether the person deserves support.
The CRPD protects access to freely chosen work in inclusive and accessible environments.3 It does not make employment a condition of dignity, care, or community membership.
Amanah must therefore hold two truths together:
A person must not prove economic usefulness to deserve care.
And:
Support needs and unconventional communication must not be used as proof that the person has nothing to contribute.
This is why Amanah needs a Participation and Possibility Map rather than a productivity score.
The continuity subject is a human life.
Not a labour unit.
Person-centred does not mean person-alone
Amanah should not turn person-centred design into the fantasy that every person must manage every part of their life without help.
Independence is not the only dignified outcome.
Interdependence is ordinary human life.
Some people will need extensive assistance.
Some will require lifelong care.
Some decisions will involve supporters, advocates, guardians, clinicians, or other lawful authorities.
Some people will choose to delegate.
Some will want a trusted person to help them communicate.
Some will need protection against coercion, exploitation, neglect, or unsafe systems.
WHO’s assistive-technology work uses a user-centred approach and describes people, policy, products, provision, and personnel as interlinked parts of access.4 The person is central, but the system around them still matters.
Similarly, NICE guidance in England and Wales states that decision support should be tailored to the individual, may include communication aids and trusted people, and should provide sufficient time, accessible information, and an environment that helps the person participate.5 That guidance is jurisdiction-specific, but it illustrates a broader design principle relevant to Amanah:
Support should increase the person’s ability to participate in decisions, not use the existence of support as evidence that their participation no longer matters.
The person-centred system is not a person-isolated system.
It is a system in which assistance remains accountable to the person it is meant to serve.
A practical architectural rule
The principle can be expressed technically:
Create one durable person principal. Attach people, roles, institutions, models, devices, records, and permissions to that principal through scoped and reviewable relationships. Never make the person a dependent attribute of another actor’s account.
This rule has several consequences.
The person persists while the care circle changes
A parent may remain.
A teacher may leave.
A clinician may change.
A centre may close.
A new advocate may be appointed.
A successor may be introduced.
The continuity remains with the person.
Authority and evidence remain different
A clinician may hold scoped decision authority.
A report has evidentiary status.
A machine observation is a hypothesis.
None of these is the person.
Expression and capture remain different
AAC, gesture, eye gaze, cards, interpreters, carers, and sensors can carry or record expression.
The expression still belongs to the person.
Contexts remain bounded
Home, school, clinical care, work, centre life, and private relationships should not collapse into one unrestricted view.
Current life can revise history
Old records cannot become permanent authority merely because they have existed longer.
The person can outlive the provider
Portability, export, succession, and vendor-failure planning are part of person-centred continuity.
This is the difference between storing a care profile and building a continuity spine.
The governing question
Amanah Companions began with an embodied-care question.
What would a companion need to know?
What should it remember?
What should it never decide alone?
How could it support communication, sensory regulation, transitions, continuity, and eventually bounded physical assistance?
Those remain important questions.
But beneath all of them is a more fundamental one:
Who is the system ultimately organised around?
If the answer is the guardian, the adult may never fully emerge from childhood.
If the answer is the institution, continuity may end at the institution’s boundary.
If the answer is the device, the life may disappear when the hardware fails.
If the answer is the company, the person may become trapped by a vendor.
If the answer is the diagnosis, the person may be reduced to a category.
If the answer is the profile, old data may begin to outrank present expression.
If the answer is the AI, assistance may quietly become artificial authority.
The answer must remain the person.
Not because the person never needs help.
Because the help is theirs.
Not because every decision can be made alone.
Because support should not erase the one being supported.
Not because a profile is useless.
Because a representation must never replace the life it represents.
The person is the continuity subject.
Everything else must know its place around them.
Research status
This essay extends the fourteen foundational Amanah Companions articles, particularly:
- What an AI Companion for a Non-Speaking Autistic Child Would Actually Need to Know
- A Parent’s, Guardian’s, or Carer’s Knowledge as Infrastructure
- Designing the Care Profile
- The Guardian Gate: Human Authority in AI-Assisted Care
- AAC Integration and the Communication Map
- Care Memory vs Surveillance Archive
- Data Projections for Continuity-Based Care AI
It also develops later authorial clarification that the supported person is the durable principal while guardians, carers, professionals, institutions, models, and devices attach through scoped roles.
This remains a research and design proposition. It does not claim that the proposed architecture has been clinically validated, legally harmonised across jurisdictions, or deployed as care software.
Sources
- United Nations, Convention on the Rights of Persons with Disabilities, especially Articles 12, 19, 21, 22, 26, and 27. https://www.un.org/development/desa/disabilities/convention-on-the-rights-of-persons-with-disabilities.html ↩
- United Nations Committee on the Rights of Persons with Disabilities, General Comment No. 1 (2014): Article 12 — Equal Recognition before the Law. https://www.ohchr.org/en/documents/general-comments-and-recommendations/general-comment-no-1-article-12-equal-recognition-1 ↩
- United Nations, Convention on the Rights of Persons with Disabilities, Article 27: Work and employment. https://www.un.org/development/desa/disabilities/convention-on-the-rights-of-persons-with-disabilities/article-27-work-and-employment.html ↩
- World Health Organization, Improving access to assistive technology, including the interlinked “5P” areas: people, policy, products, provision, and personnel. https://www.who.int/activities/improving-access-to-assistive-technology ↩
- National Institute for Health and Care Excellence, Decision-making and mental capacity (NG108) and Quality statement 1: Supported decision making (QS194). These sources apply specifically within the England and Wales legal and service context but are cited here for their practical treatment of communication support, decision-specific participation, accessible information, time, and trusted support. https://www.nice.org.uk/guidance/ng108 and https://www.nice.org.uk/guidance/qs194/chapter/Quality-statement-1-Supported-decision-making ↩
© 2026 • MITHAQ PRAXIS • CC BY-NC-ND 4.0 Unless Otherwise Stated.