A tightly governed assistive system may begin with an apparently reasonable rule:

Only formally diagnosed people can use it.

The logic is easy to understand.

A diagnosis creates a recognised entry point.

A registered professional can confirm eligibility.

A clinic, school, centre, or government service can attach the person to an existing system.

Sensitive functions can be restricted to verified users.

Professional recommendations can be signed.

The technology is less likely to be purchased casually, configured recklessly, or used outside the population it was designed to support.

For high-risk embodied systems, some form of professional assessment, regulated provision, or formal eligibility may eventually be necessary.

But diagnosis is not distributed equally.

It depends on whether a family can afford assessment.

Whether trained professionals exist nearby.

Whether a school recognises the signs.

Whether a healthcare system has a pathway.

Whether the waiting list is measured in weeks or years.

Whether the person’s presentation matches the patterns professionals were trained to recognise.

Whether the family speaks the dominant language.

Whether disability is understood, minimised, hidden, feared, or stigmatised in the surrounding culture.

Whether an adult who learned to mask can persuade anyone that the struggle is real.

If Amanah requires formal diagnosis for every form of access, it will inherit those inequalities.

The gate may become safer.

It may also become a wall.

Diagnosis is not equally available

Autism is identified through systems, not in a vacuum.

The World Health Organization states that autistic characteristics may be detected in early childhood, but diagnosis often occurs much later. It also notes that prevalence in many low- and middle-income countries remains unknown and that autistic people experience unmet healthcare needs, stigma, discrimination, and barriers to participation.¹

The problem is larger than autism alone.

The WHO–UNICEF Global Report on Children with Developmental Disabilities describes children and young people with developmental disabilities as neglected in health-system planning and policy, while continuing to experience stigma, institutionalisation, barriers to healthcare, and inequalities in health and education.²

A systematic review comparing autism diagnostic pathways across countries of different income levels found an average delay of more than thirty-two months between families’ first concerns and diagnosis across the studies it reviewed. The authors also highlighted the limited research available from low- and middle-income countries and the need for better recognition and diagnostic infrastructure.³

Even within wealthy countries, access is unequal.

A 2026 systematic review of diagnostic inequities identified interacting barriers including limited health literacy, insurance and financial constraints, provider bias, organisational failures, culturally insensitive assessment tools, stigma, medical jargon, and inadequate language interpretation.⁴

This means “diagnosed” and “undiagnosed” are not neutral categories.

They are partly outcomes of the systems surrounding the person.

The same need can receive different recognition

Imagine two children with similar communication and sensory needs.

One lives in a city with paediatric specialists, early screening, trained teachers, insurance or public funding, and a centre capable of coordinating assessment.

The other lives in a place where the nearest specialist is far away, the family cannot afford private assessment, the school calls it a speech delay, and relatives insist the child will grow out of it.

The first child may receive:

• a diagnosis;
• AAC assessment;
• school accommodations;
• therapy;
• disability registration;
• assistive technology;
• and access to future Amanah services.

The second child may have many of the same needs and receive none of them.

A diagnosis-gated Amanah system would not merely respond to inequality.

It would convert the inequality into product eligibility.

The same problem appears in adulthood.

One adult receives a late diagnosis after years of research, referrals, and expense.

Another recognises similar patterns but cannot access assessment.

A third lives in a culture where pursuing diagnosis could produce family conflict, employment risk, stigma, or social exclusion.

A fourth is repeatedly assessed for anxiety, depression, personality, or behaviour without anyone considering autism or another neurodevelopmental condition.

A fifth needs communication, executive-function, sensory, or transition support but may never meet criteria for one particular diagnosis.

The gate must be able to see need without pretending to diagnose.

Culture is not a decorative variable

Technology projects often mention culture near the end.

Translate the interface.

Change the illustrations.

Consult the community.

But culture affects whether the person reaches the gate at all.

It shapes:

• which differences are noticed;
• which are tolerated;
• which are punished;
• which are medicalised;
• who is blamed;
• whether families seek help;
• which professionals are trusted;
• whether disability is discussed publicly;
• whether a child is hidden;
• whether an adult is expected to keep coping;
• and whether formal records are seen as protection or danger.

UNICEF describes disability stigma as rooted in social and cultural beliefs that may devalue the lives of disabled children and young people.⁵ Its broader disability work also identifies stigma, discrimination, inaccessible services, and weak disability-inclusive systems as barriers to essential support.⁶

Culture can delay diagnosis.

It can also protect people in ways formal systems do not.

Extended family, religious communities, neighbourhood networks, and local carers may carry practical knowledge that institutions overlook.

Amanah should not treat every non-clinical system as ignorance.

Nor should it romanticise family or community structures that deny, hide, or control the disabled person.

The architecture needs to preserve cultural context without allowing culture to become an excuse for withholding communication, healthcare, education, or personhood.

Financial access is part of diagnostic access

A diagnosis may require:

• consultation fees;
• travel;
• time away from work;
• school reports;
• repeated appointments;
• specialist assessments;
• language interpretation;
• private therapy while waiting;
• and the ability to navigate forms, referrals, and institutions.

A family may technically live in a country with diagnostic services and still be unable to use them.

The same is true for assistive technology.

WHO estimates that more than 2.5 billion people need one or more assistive products, yet access can be as low as 3 percent in some settings, compared with as high as 90 percent in some wealthier settings. Reported barriers include cost, low awareness, physical inaccessibility, limited product choice, weak procurement, workforce shortages, inadequate policy, insufficient funding, and fragmented provision.⁷

Amanah cannot solve all of these structural failures.

But it must not pretend that a formal diagnosis makes technology accessible.

A diagnosed person may still lack:

• internet access;
• an AAC device;
• trained communication partners;
• a compatible phone;
• private space;
• transport;
• professional support;
• or the money required for ongoing service, repair, and replacement.

Eligibility and access are not the same thing.

Registration can protect—and expose

The idea of registered access has real value.

A professional identity could allow Amanah to verify that:

• a person is licensed within a recognised profession;
• the credential is current;
• the professional is acting within scope;
• a recommendation was genuinely submitted by them;
• and the authority expires when the relationship or credential ends.

A disability registration or service record may also allow access to public funding, centre provision, subsidies, specialised equipment, or legally recognised accommodations.

But registration carries risk.

A central disability identifier can expose sensitive information.

A national system may be efficient while providing weak control over who can see what.

A family may fear that registration will affect education, employment, insurance, migration, marriage, custody, or social standing.

A person may want access to support without disclosing diagnosis to every device, school employee, transport worker, or service provider.

Amanah should therefore distinguish:

• identity verification;
• professional credential verification;
• disability or service eligibility;
• diagnosis;
• functional need;
• and permission for a particular action.

These are not the same check.

The system may need to confirm:

This is the correct person.

It may also need to confirm:

This therapist holds a current credential.

It does not follow that every participant needs to see:

This person has this diagnosis.

The gate should verify only what the function requires.

Digital identity is not the same as care authority

A strong national identity system may make Amanah easier to integrate.

It can reduce duplicate records.

It can verify professionals.

It can connect public services.

It can simplify eligibility, funding, and portability.

But digital identity answers:

Who is this?

It does not automatically answer:

What may this person do?

Nor:

What does the supported person want?

Nor:

Which information is relevant?

Nor:

Does a current diagnosis exist?

Nor:

Should a robot be allowed to act?

Amanah needs several separate layers:

1. identity — who the actor is;
2. credential — which role or qualification they hold;
3. relationship — how they are connected to the supported person;
4. eligibility — which service or capability is available;
5. authority — what they may decide or approve;
6. permission — what they may view or do now;
7. person participation — what the supported person expressed;
8. audit — what actually happened.

A country can have excellent digital identity and poor person-centred care.

It can also have fragmented technology and strong disability rights.

Amanah readiness cannot be measured by digital integration alone.

A diagnosis should be one evidence pathway

The solution is not to make diagnosis irrelevant.

It is to stop treating diagnosis as the only legitimate evidence of need.

Amanah may eventually recognise several pathways.

Formal diagnosis

A diagnosis from a qualified professional under the applicable national system.

This may be required for:

• statutory disability benefits;
• certain clinical pathways;
• publicly funded services;
• regulated medical claims;
• or high-risk interventions.

Provisional or working formulation

A qualified professional may believe that a condition is likely while assessment remains incomplete.

This should not be represented as confirmed diagnosis.

But it may justify interim support.

Functional assessment

A professional or multidisciplinary team documents communication, sensory, motor, executive, daily-living, or accessibility needs without requiring one diagnostic label to explain all of them.

Education or centre support assessment

A school, disability centre, or care service identifies accommodations required in that setting.

Its authority should remain limited to the relevant context.

AAC or assistive-technology assessment

A person may have a demonstrated communication or access need regardless of whether a broader diagnosis is confirmed.

Person-reported need

An adult may identify sensory, communication, executive-function, or transition support needs.

For low-risk tools, the person’s own account should matter.

Caregiver observation

A parent or carer may document recurring difficulties or supports.

This is not diagnosis.

It is still useful evidence.

Machine-detected pattern

An AI system may identify something worth reviewing.

This is the weakest path for eligibility and should never become diagnosis or automatic authority.

These pathways should remain distinct.

A caregiver observation should not impersonate a clinical assessment.

A self-report should not be relabelled as a professional finding.

A professional recommendation should not become permanent fact outside its scope.

But the absence of a diagnosis should not erase every other form of evidence.

The gate should ask what is being unlocked

The fairest access rule is not:

Diagnosis required: yes or no.

It is:

What capability is being requested, and what evidence is proportionate to its risk?

A visual planning tool is not the same as a medication function.

A person-owned sensory journal is not the same as clinical monitoring.

An AAC communication passport is not the same as an autonomous physical assistant.

A reflective companion is not the same as a system making decisions during intimate care.

A reminder is not the same as restraint.

Amanah therefore needs a layered gateway.

The full risk architecture belongs in the next article.

But the governing principle begins here:

Low-risk support should not be withheld merely because formal diagnosis is unavailable. Higher-risk functions may require stronger evidence, professional recommendation, regulated provision, training, and oversight.

This protects both safety and access.

Open support without diagnostic claims

Some Amanah functions could be available without formal diagnosis, provided they remain bounded and do not make clinical claims.

Examples may include:

• a personal communication passport;
• visual schedules;
• transition planning;
• sensory journalling;
• person-owned preferences;
• routine reminders;
• accessible preparation for appointments;
• a record of what helps and what hurts;
• private reflection;
• executive-function support;
• AAC-adjacent organisation;
• support for documenting experiences before an assessment;
• and exportable summaries the person chooses to share.

These tools may help autistic people.

They may also help people with ADHD, developmental disabilities, acquired disabilities, anxiety, trauma, chronic illness, cognitive differences, communication difficulties, or no diagnosis at all.

The tool does not need to decide why a visual schedule helps.

It needs to help safely.

It must not advertise itself as diagnostic treatment.

It must not infer that every user is autistic.

It must not silently construct a diagnosis profile.

It must not route data into model training.

It must not pressure the person to seek diagnosis simply to unlock ordinary support.

Supported access without a final diagnosis

A second layer may require involvement from a professional, school, centre, or recognised service without demanding a final diagnosis.

This matters when:

• assessment is pending;
• the person is on a long waiting list;
• a child has clear developmental or communication needs;
• a professional can document functional need;
• a school requires accommodations now;
• or a family needs interim support.

WHO’s Caregiver Skills Training programme was designed for caregivers of children with developmental delays or disabilities, including autism, and has been used across diverse settings. Its framing is relevant because it does not require families to wait for a perfect specialist system before receiving practical support.⁸

Amanah could adopt a similar principle for selected functions:

Support the demonstrated need while preserving diagnostic uncertainty.

That is not lowering the standard.

It is refusing to make delay itself part of the harm.

The system should not reward diagnosis inflation

A diagnosis-gated product can create a perverse incentive.

People may pursue a particular label not because it is the most accurate clinical explanation, but because the label unlocks technology, funding, or services.

Professionals may face pressure to certify eligibility.

Companies may market assessment pathways tied to product access.

Families may feel that without the right code, their child will be left behind.

Amanah should avoid creating:

Get diagnosed to unlock the good version.

That does not mean diagnosis should never affect access.

It means access design should not encourage medicalisation beyond what the function requires.

A person should not need a lifelong identity label to use a low-risk planning tool.

A family should not need to prove severity every year to retain a communication passport.

A professional should not be asked to diagnose merely to authorise an accommodation they can justify functionally.

The gate must be difficult to exploit without forcing every need through one clinical bottleneck.

Amanah could help the path to assessment

Amanah should not diagnose.

It may still improve the assessment process.

With the person’s consent and careful governance, it could help organise:

• developmental history;
• communication patterns;
• sensory experiences;
• periods of speech loss or shutdown;
• school observations;
• caregiver concerns;
• previous assessments;
• accommodations tried;
• supports that helped;
• contexts where difficulties appear;
• and the person’s own account.

A clinician often sees only a short appointment.

A structured longitudinal record may help preserve information that would otherwise be forgotten, scattered, or reconstructed under stress.

But the export must retain provenance.

It should distinguish:

• person report;
• caregiver observation;
• school note;
• clinical record;
• and machine-generated summary.

The output should say:

This document supports assessment. It is not a diagnosis.

Amanah can shorten the distance to understanding without pretending to be the final authority.

Undiagnosed does not mean ungoverned

Opening low-risk access does not require abandoning safeguards.

An undiagnosed user should still receive:

• privacy protections;
• source trace;
• clear limits;
• transparent AI labelling;
• export control;
• deletion options;
• consent and assent safeguards;
• no-training rules;
• age-appropriate design;
• and escalation when the system encounters risk it cannot manage.

A diagnosis does not create these protections.

The architecture does.

Similarly, an undiagnosed person should not receive a high-risk capability simply because they self-identify.

The capability must meet the correct safety and authority requirements.

The gate should neither say:

no diagnosis, no support

nor:

self-identification, therefore unlimited access

It should say:

What support is requested? What is the risk? What evidence is available? Who is responsible? What may the system safely do?

Public systems must not become the only systems

Government registration and centre-based provision may be appropriate for some Amanah layers.

They can provide:

• professional oversight;
• public funding;
• equipment maintenance;
• incident reporting;
• credential verification;
• procurement standards;
• and continuity across services.

But not every country has a reliable integrated system.

Even where systems exist, people fall outside them.

Migrants.

Rural families.

People without documentation.

Adults who were never assessed.

People whose records were lost.

People who do not trust the state.

People living under stigma.

People moving across borders.

People whose diagnosis is recognised in one place but not another.

Amanah therefore needs both:

• institutional pathways; and
• person-portable pathways.

The continuity spine should not disappear when somebody leaves the system that first registered them.

Global inequality must be designed in from the beginning

Amanah will not be implemented everywhere at the same time.

Some countries have:

• national identity infrastructure;
• integrated health records;
• established disability services;
• professional registries;
• assistive-technology programmes;
• robotics laboratories;
• public procurement;
• and strong digital regulation.

Others rely on fragmented clinics, charities, family care, paper records, community organisations, or privately funded services.

It would be easy to build Amanah in a highly digitised country and call the architecture universal.

It would not be universal.

The global implementation map planned for this series must therefore assess more than technological readiness.

It should examine:

• diagnosis accessibility;
• professional workforce;
• disability registration;
• cultural stigma;
• AAC provision;
• privacy and data law;
• supported decision-making;
• centre capacity;
• public funding;
• offline infrastructure;
• language diversity;
• rural access;
• and whether undiagnosed people have any legitimate support pathway.

A system developed only where records are clean may fail precisely where continuity is most fragile.

The gate must protect without humiliating

There is also a human experience to consider.

Every eligibility process asks the person or family to prove something.

Prove the diagnosis.

Prove the disability.

Prove the severity.

Prove the financial need.

Prove that support is still necessary.

Prove that the person cannot simply cope.

This can become exhausting.

Amanah should minimise repeated proof.

Once legitimate evidence is recorded, the system should preserve it under the person’s control and allow purpose-limited reuse where lawful.

It should not require the family to retell the entire life at every doorway.

It should not make the person perform distress for the system.

It should not force an adult to expose intimate history to access a simple accommodation.

It should not design verification as suspicion.

The gate should be firm about authority and gentle with the person.

The wrong gate

The wrong Amanah gate says:

Diagnosis first. Support later.

It assumes diagnosis is equally reachable.

It treats one label as the truth of the person.

It makes professionals universal gatekeepers.

It gives registered systems priority over lived need.

It may exclude exactly the people most affected by poverty, geography, stigma, culture, language, or weak institutions.

It creates a premium category of recognised disability and a shadow category of unsupported need.

The better gate

The better gate asks:

What does this person need now?

What capability is being requested?

How risky is it?

What evidence supports the need?

Which evidence path is legitimate for this function?

Does a qualified professional need to be involved?

What is the person’s role?

Who remains accountable?

What information is actually necessary?

When does the permission expire?

Can the person refuse?

Can the support travel if the institution changes?

Diagnosis remains one important answer.

It is not the only answer.

The governing sentence

Amanah must not make safety and equity enemies.

A fully open system could expose vulnerable people to poorly governed clinical or embodied functions.

A diagnosis-only system could deny useful support to people who cannot reach the diagnostic gate.

The architecture must hold both truths.

The gate must protect the person without becoming another wall against them.

Formal diagnosis may be necessary for some services.

Professional recommendation may be appropriate for others.

Functional need may be enough for bounded support.

Self-directed low-risk tools may remain open.

The system should preserve the difference.

Because a person’s need does not begin on the day the paperwork arrives.

And dignity should not depend on living in the right country, meeting the right professional, speaking the right language, having enough money, or presenting disability in the form a system already knows how to recognise.

---

Research status

This essay extends the fourteen foundational Amanah Companions articles, particularly:

• Designing the Care Profile
• The Guardian Gate: Human Authority in AI-Assisted Care
• Human-Led Care in an AI-Assisted Future
• AAC Integration and the Communication Map
• Data Projections for Continuity-Based Care AI
• AI Prototype Architecture for an Amanah Companion System

It also incorporates later authorial clarification concerning diagnosed and undiagnosed users, cultural and financial barriers to assessment, professional identity, government systems, and the danger of allowing formal diagnosis to become the universal gate to assistive technology.

This is a design and research proposition. Eligibility for disability services, clinical technologies, educational accommodations, public funding, medical devices, and embodied assistance remains jurisdiction-specific. Amanah should not diagnose users or represent non-clinical evidence as equivalent to a formal diagnosis.

Sources

1. World Health Organization, Autism, 17 September 2025. WHO notes that autism is often diagnosed later than early childhood, that abilities and needs vary and evolve, that prevalence in many low- and middle-income countries remains unknown, and that autistic people face stigma, discrimination, and unmet healthcare needs. https://www.who.int/news-room/fact-sheets/detail/autism-spectrum-disorders ↩
2. World Health Organization and UNICEF, Global Report on Children with Developmental Disabilities: From the Margins to the Mainstream, 2023. https://www.who.int/publications/i/item/9789240080232 ↩
3. Luísa M. B. Rocha et al., “Autism Spectrum Disorder Diagnoses: A Comparison of Countries with Different Income Levels,” International Journal of Environmental Research and Public Health 19, no. 15 (2022): 9268. https://doi.org/10.3390/ijerph19159268 ↩
4. Komal Bajaj et al., “Bridging Diagnostic Safety and Mental Health: A Systematic Review Highlighting Inequities in Autism Spectrum Disorder Diagnosis,” 2026. The review identified interacting U.S.-based barriers including health literacy, insurance, provider bias, organisational structure, culturally insensitive tools, jargon, and inadequate interpretation. Its findings should not be treated as a direct global prevalence estimate. https://pmc.ncbi.nlm.nih.gov/articles/PMC12707049/ ↩
5. UNICEF, Disability Inclusion Frontline Workers Training: Focus Brief — Stigma and Discrimination. https://www.unicef.org/documents/disability-inclusion-frontline-workers-training-focus-brief-stigma-and-discrimination ↩
6. UNICEF, Children with Disabilities. https://www.unicef.org/disabilities ↩
7. World Health Organization, Assistive Technology. WHO reports access as low as 3 percent in some low-income settings and identifies barriers including cost, awareness, physical access, product availability, procurement, workforce capacity, policy, funding, fragmentation, and sociodemographic obstacles. https://www.who.int/news-room/fact-sheets/detail/assistive-technology ↩
8. World Health Organization, “WHO’s Training for Caregivers of Children with Autism Goes Online,” 31 March 2022. The Caregiver Skills Training programme serves caregivers of children with developmental delays or disabilities, including autism, and has been piloted across diverse countries and community settings. https://www.who.int/news/item/31-03-2022-who-s-training-for-caregivers-of-children-with-autism-goes-online ↩