HSIMC · UN Six-Language White Paper · P6 / W0-017

Part I - The Opportunity: Bringing Intelligence into Everyday Making

Part I at a glance

Generative AI can produce plans, code, drawings, analyses and alternatives at a speed that is increasingly difficult for the physical world to match. A material still has to be found. A machine still has to be available. A fixture has to fit. A person must have authority to act. A result must be accepted, repaired, handed over, paid for or rejected. The opportunity described in this white paper is to make those physical capabilities easier to discover, combine and verify without pretending that the real world has become frictionless.

Part I establishes three propositions.

First, the distance between an answer and an outcome is not a temporary inconvenience. It is a design space. Engineering, manufacturing and service work become more valuable when AI can work with explicit physical constraints, responsible parties and evidence rather than with abstract plans alone.

Second, a familiar credit-card-sized footprint can serve as a practical starting point for participation. It does not mean that every function must be compressed into one card. It means that a small, portable and understandable entry can connect printed matter, NFC or QR, local electronics, a computing node, an existing machine and cloud services. Intelligence can be distributed across layers while the interface remains understandable.

Third, broad participation requires more than access to technology. It requires authority, meaningful contribution and a fair way to recognize value. The United Nations promise to leave no one behind provides an important human principle: identify who is excluded, understand why, enable meaningful participation and keep accountability visible. HSIMC applies that principle to a new question for the GenAI physical era: who is able to help turn plans into useful capability, under what conditions, and who shares in the value created?

The chapters that follow are therefore not a product specification. They are the foundation for a collaboration model that will be tested through materials, making, engineering, logistics and real-world execution.


Chapter 1 - The Distance Between an Answer and an Outcome

1.1 The physical work behind a digital answer

A model can answer a question in seconds. It can propose a circuit, create code, compare materials, generate a drawing, suggest a process plan or produce a set of tests. Yet none of these outputs is, by itself, a physical result.

A useful physical result still depends on conditions that are often missing from the prompt: - the actual material grade, thickness, moisture, finish or damage; - the real machine, controller, firmware and tooling available at a location; - the dimensions and tolerances of a fixture or connector; - the skill and availability of the person who will perform the operation; - the space, access window and safety rules at the site; - the authority to use the resource, modify the object or move the asset; - the current status of inventory, power, network or transport; - the method by which the result will be inspected and accepted.

This distance between an answer and an outcome is easy to underestimate because digital work hides much of the physical effort. A PDF can contain thousands of pages before anyone has handled a sheet of paper. Thousands of QR codes can be generated before anyone has cut, laminated, punched or attached a card. A PCB can be drawn before a fabricator has confirmed stack-up, copper weight, drill limits or component availability. A warehouse plan can be optimized before anyone has confirmed that the aisle is clear or the pallet is structurally sound.

The gap is not evidence that AI is failing. It is evidence that the next engineering problem is different.

The problem is no longer only, “Can a model produce a plausible plan?” It is increasingly, “Can a plan be connected to real resources, real authority, real constraints and a verifiable outcome?”

That is the opportunity HSIMC addresses.

A result is more than an action

This white paper uses a deliberately concrete unit of work:

A defined result, at a named place and time, under agreed conditions, completed by a responsible party and verified by evidence.

A result might be: - a sample cut to an agreed geometry; - a printed marker placed within a defined alignment tolerance; - a conductive textile pattern whose electrical properties are measured; - a material transferred from a storage location to a machine; - a finished item placed into an identified temporary location; - a baggage handover completed within an accepted time window; - an empty trolley moved away from a constrained passenger area; - an inspection completed and recorded against a versioned method.

The definition is intentionally neutral about whether the work is performed by a person, a machine, a robot, a service provider or a combination of them. Automation is one possible capability. It is not the definition of value.

The value lies in the accepted change of state.

1.2 The November engineering invitation

On 3 November 2026, OpenAI published Accelerating engineering for the physical world, describing advances in how AI models can support mechanical design, electrical engineering, software, simulation and end-to-end system design. The publication emphasizes that engineers must bring together specialized tools, manufacturing requirements and changes across disciplines, and it invites engineers, engineering organizations and software vendors to help shape the work around real problems and real tools.

HSIMC treats that publication as an important engineering signal, not as an endorsement of HSIMC or of the proposals in this white paper.

The complementary opportunity is clear. Frontier models can become increasingly capable at design and analysis, while many of the resources needed to make a result real remain local, fragmented and difficult to describe in a form another team can safely use.

A small workshop may have: - a laser cutter but no standardized capability profile; - a sewing or embroidery machine but no agreed electronic test method; - a printer capable of producing a functional mark but no versioned alignment procedure; - a CNC machine whose operator knows the true limits better than its public brochure; - a box of reusable labels or materials whose current state is known only to one person; - a useful jig that exists only as a physical object; - a reliable local supplier that does not expose an API.

A global engineering team may have excellent code, simulation or analysis but no fast path to those capabilities.

The opportunity is therefore not only to make AI better at engineering. It is also to make engineering capability more legible to AI and to other people.

From hidden capability to a usable contribution

For a capability to become useful across teams, at least five questions must be answerable:

  1. What can it do?
    The function, input, output and operating conditions.

  2. Where and when is it available?
    The place, time window, preparation requirements and current status.

  3. Who is responsible?
    The person or organization that can accept the work and respond when conditions change.

  4. Who has authority?
    The party allowed to view data, reserve capacity, operate equipment, change settings, release results or cancel work.

  5. How will the result be checked?
    The evidence, test, measurement, inspection or handover that turns an activity into an accepted outcome.

These questions are as relevant to a person with a sewing machine as they are to a large automated factory.

This is why HSIMC’s working model is built around Resource, Authority and Evidence. - Resource describes what exists and under what conditions it may be useful. - Authority describes who may decide, operate, share, accept or withdraw. - Evidence describes what was actually observed, tested, delivered or rejected.

The model is deliberately compatible with imperfect reality. A resource can be registered but unavailable. A machine can be capable in principle but unverified in a particular configuration. A file can exist while its physical implementation remains untested. A human expert can be available only during a narrow time window. A local supplier can contribute through a structured handoff even if it has no sophisticated software stack.

That flexibility matters because much of the physical world does not begin as a clean digital twin.

1.3 The first design principle: make the next step executable

The most useful AI output is often not a complete master plan. It is the next step that another party can actually perform.

Consider a simple example: a printed sample must be aligned and inspected.

A global model can propose an alignment pattern. A local printer can produce the sample. A fixture maker can define the mechanical reference. A vision engineer can write a locate_marks function. A camera integrator can connect a real device. A second workshop can reproduce the test. None of those contributors needs to own the entire system.

The collaboration becomes valuable when their outputs meet through a common contract: - known input image or real camera frame; - known pattern version; - coordinate system and units; - expected output fields; - acceptable error cases; - calibration status; - software version; - evidence file or receipt.

The first workshop may deliver only the printed target and fixture. That is enough if the contribution is understandable and reproducible.

This is the deeper meaning of the project invitation:

Bring your existing capability. Deliver one reproducible function.

It lowers the threshold for useful participation while increasing the quality of the handoff.

Why “now” matters

The physical world has always contained distributed capability. What changes in the GenAI era is the speed at which design, explanation, code, translation and coordination can be produced around that capability.

A local workshop no longer needs to prepare a complete international product catalogue before a remote engineer can understand its contribution. A model can help translate technical notes, generate interface examples, structure a test plan and compare alternatives. A remote team can begin with a simulator while hardware is still being prepared. Evidence can be returned in a form another team can inspect.

This does not eliminate local work. It raises the value of local work that is clearly described.

The competitive variable begins to shift from “Who can manufacture the cheapest unit?” toward a richer question:

“Who can provide the right capability, at the right place and time, with a reliable handoff and rapid iteration?”

Low cost remains important in many markets. But for prototypes, repairs, customization, field integration, urgent replacement, small batches and local service, speed and coordination can be economically decisive.

This is especially visible in dense industrial environments, where mature local facilities may already have printing, laser processing, assembly, packaging and repair capabilities. The opportunity is to connect them to new forms of engineering demand without requiring each participant to become a full-stack technology company.

1.4 What this white paper does not claim

A credible physical-world initiative must say what is not yet proven.

This white paper does not claim that: - every local workshop can produce electronics safely; - every recycled material is environmentally preferable; - every sewing or printing process can become an electronic manufacturing process; - every QR or NFC interaction authorizes physical control; - UWB positioning automatically makes motion safe; - a simulator result proves the behavior of a real device; - a successful cargo movement validates a passenger-carrying system; - a credit-card-sized concept is already a complete card-thickness computer; - a global network exists simply because a website lists potential resources.

Instead, it proposes a method for moving from proposed -> documented -> simulated -> measured -> independently reproduced -> commercially accepted, while preserving the evidence level at each step.

That evidence discipline is essential to the opportunity. It allows ambition and engineering credibility to coexist.


Chapter 2 - A Credit-Card-Sized Starting Point

2.1 A familiar footprint and an explicit design target

A standard credit card is one of the most familiar engineered objects in daily life. The ISO/IEC 7810 ID-1 format is nominally 85.60 mm by 53.98 mm. The full standard also addresses physical characteristics beyond planar size, so this white paper uses the footprint as a design reference, not as a claim that every proposed device already meets the complete card standard.

Why begin there?

Because a familiar physical scale is powerful.

People can immediately imagine: - carrying it; - attaching it to another object; - placing it in a wallet, package or fixture; - printing on it; - adding a QR code; - integrating an NFC inlay; - stacking it with other layers; - using it as an identity and service entry point.

A credit-card-sized starting point turns an abstract infrastructure discussion into something a person can hold.

But the design principle is not “put everything inside one card.”

It is “make the entry small enough to understand, while allowing the capability behind it to be distributed.”

That distinction prevents the form factor from becoming a gimmick.

2.2 The card is an entry, not the whole machine

The physical stack can be understood as several layers.

Layer A - the visible entry

This may be a printed card, label, tag, patch, plate or textile surface. It can carry: - human-readable identity; - QR entry; - NFC entry; - alignment marks; - instructions; - a product or resource reference; - a version or batch marker.

It can be manufactured with mature local processes and does not need to contain the full computing system.

Layer B - the functional surface or interface

This may add: - conductive thread; - printed electrodes; - a simple antenna; - a flexible conductor; - a sensor contact; - a connector or breakout; - a mechanical locating feature.

The purpose is not to romanticize craft. The purpose is to determine whether a familiar process can deliver a measurable function.

For example, conductive embroidery can form an antenna geometry, but the result still requires a compatible chip or module, electrical matching and testing. Printing can produce a marker or electrode, but its electrical or optical performance still depends on material, geometry and process control.

The value of the maker’s contribution becomes clearer when the function is separated from the rest of the system.

Layer C - the electronic module

Existing electronic modules can provide functions that a local maker should not need to reinvent: - NFC or RFID interface; - UWB ranging or positioning; - camera or sensor input; - USB or serial conversion; - secure storage; - wireless connectivity; - power conditioning.

A good architecture makes these modules replaceable and testable rather than permanently entangling them with a single craft process.

Layer D - the local computing node

The current HSIMC engineering route uses the T736 family as the shortest reference path for a portable computing and physical-interface node. The existing project evidence supports a documentary hardware baseline and a simulator-based engineering route; it does not yet prove a finished credit-card-sized product.

The intended role of the node is broader than “a phone.” It may operate as: - a portable team node; - a desktop local server; - an embedded interface node; - a gateway between existing equipment and cloud services; - a local evidence collector; - a runtime for adapters and constrained functions.

Its value comes from the common interface and engineering workflow, not from the label attached to the enclosure.

Layer E - existing equipment and cloud capability

The card or local node may connect to: - an existing printer; - a laser machine; - a sewing or embroidery machine; - a camera; - a barcode or RFID reader; - a UWB module; - a machine controller; - a warehouse location; - a remote engineering workspace; - a model or service in the cloud.

This is where a small physical entry can unlock a much larger capability.

2.3 Why distributed intelligence matters

The physical world is full of mature machines that were never designed for modern GenAI workflows.

Replacing all of them would be economically and environmentally unrealistic.

A more practical strategy is to separate: - the stable physical capability; - the replaceable interface; - the local compute and control boundary; - the remote model or workflow; - the human authority to use them.

This allows new intelligence to be added incrementally.

A 20-year-old machine may continue performing the motion it already performs well. A new adapter can expose a small set of safe operations. A T736-class node can translate data, hold local state or connect a camera. A remote model can help generate settings or inspect results. A human operator remains responsible for the site-specific authority and exception handling.

The architecture therefore treats existing things as potential capability, not as obsolete infrastructure waiting to be replaced.

This is particularly important for small enterprises and local makers. Their competitive advantage is often not a proprietary digital platform. It is proximity, practical experience, equipment access, responsiveness and the ability to improvise within known constraints.

The challenge is to turn that tacit capability into something another team can rely on.

2.4 From stone to a useful component

“Stone” in the title is not a claim that construction waste can directly become electronics. It represents the most basic question of local making:

What useful function can an available material perform, after we understand its source, properties and processing conditions?

A material stream from construction or manufacturing may contain mineral, polymer, wood, metal or mixed fractions. Each requires its own characterization and appropriate processing route. A responsible project does not label an item “recycled” and assume the environmental case is complete.

The useful workflow is:

  1. identify the source and composition as far as practical;
  2. define the intended function;
  3. process a sample using a versioned method;
  4. measure relevant performance;
  5. record yield, waste, energy, labour and cost;
  6. compare against the baseline alternative;
  7. decide whether the material is suitable for that function.

A material may be useful as: - a fixture body; - a non-electrical structural plate; - a display or information surface; - a protective component; - a sample substrate; - packaging; - a temporary jig; - a composite ingredient.

The important point is not the number of possible uses. It is that each use can become an explicit capability with evidence.

That is how an “ordinary” material becomes part of an intelligent production network.

2.5 From thread to a functional layer

“Thread” represents a different path: the transformation of a familiar human skill into a measurable technical contribution.

Sewing and embroidery are highly distributed capabilities. They exist in large factories, small workshops, schools, homes and community settings. Conductive thread makes it possible to explore electrical patterns using equipment and skills that are already widely available.

The contribution might be: - an embroidered antenna geometry; - a conductive path; - a wearable contact; - a stitched sensor interface; - a repeatable placement method for a small electronic module.

But the white paper deliberately separates making the pattern from proving the electronic performance.

A maker may be able to reproduce stitch density and geometry with high consistency. An RF engineer may be needed to characterize impedance and antenna performance. An electronics partner may supply the compatible chip. Another team may test durability after bending or washing.

This is exactly the kind of multi-party engineering that GenAI can help coordinate.

The network does not require one person to know everything. It requires a shared definition of the result and a reliable handoff between contributors.

2.6 The card as a contract with the physical world

The most important function of the card may ultimately be neither computing nor identification.

It may be coordination.

A physical card or tag can point to a persistent resource identity. From there, an authorized user or agent may learn: - what the object is; - who is responsible for it; - which capabilities are available; - which documents are current; - what evidence exists; - what actions can be requested; - what permissions are required; - what work is already in progress.

The physical entry is simple. The responsibility behind it is explicit.

This is why QR and NFC are useful but not sufficient. They solve discovery and proximity. They do not, by themselves, solve authority, freshness, safety or accountability.

The card therefore becomes the front door to a larger capability, not the capability itself.


Chapter 3 - Leaving No One Behind in the GenAI Physical Era

3.1 A human principle for a technical transition

The United Nations 2030 Agenda describes 17 Sustainable Development Goals and 169 targets as integrated and indivisible, and includes the pledge that no one will be left behind and that the furthest behind should be reached first. United Nations development guidance further frames Leave No One Behind as a central transformative promise, emphasizing the need to identify who is excluded and why, enable meaningful participation and maintain accountability.

This white paper does not claim that HSIMC represents the United Nations or that every project activity advances every SDG.

Instead, it adopts a practical question inspired by that principle:

As GenAI enters the physical economy, who gains the ability to participate - and who may be left outside the new production system?

The question matters because technological transitions can concentrate capability even while lowering the cost of information.

A person may have access to a powerful AI model but still lack: - a machine; - a safe workplace; - a recognized identity in a supply network; - capital for tooling; - language access; - reliable connectivity; - a way to prove quality; - a channel to receive paid work; - authority over the data or equipment involved.

Conversely, another person may have valuable physical capability but no easy way to expose it to a global engineering workflow.

A seamstress may understand repeatable textile production but never describe the capability as an engineering service. A retired technician may know a machine better than its documentation. A small print shop may be able to produce a same-day functional sample that is more valuable to an engineer than a cheaper overseas batch arriving weeks later. A local operator may know that a pallet, corridor or fixture is unsafe even when the digital plan says otherwise.

These people are not “non-technical.” They hold forms of capability that the digital system often fails to recognize.

3.2 Participation is not the same as access

Giving someone a login does not create meaningful participation.

Meaningful participation requires at least four things.

1. The person can understand the work

The task must be expressed in a language and form appropriate to the participant. Technical detail should be available when needed, but the first interface should make the expected result understandable.

This is one reason the project uses layered publication: card, brief, summary and full white paper. Different people need different depths of explanation.

The same principle applies to engineering work. A machine operator needs the acceptance condition for the operation. A hardware engineer needs electrical limits. A software engineer needs the function contract. A customer needs to know the result, timing, price and responsibility.

2. The person can decide whether to participate

Authority must remain visible.

A resource being discoverable does not imply that it may be used. A person being listed does not imply that they have accepted a job. A machine being technically capable does not imply that remote control is permitted. A document being readable does not imply authority to change the physical system.

This is why HSIMC separates Resource from Authority.

3. The contribution can be recognized and verified

A contribution should leave evidence that can be attributed to the responsible party and version.

That evidence may be: - a sample; - a measurement; - a photograph; - a machine log; - a code commit; - a test report; - a handover receipt; - an inspection result; - a signed acceptance; - a reproducible procedure.

Recognition matters because invisible work is easily undervalued.

4. The participant can share in the value

Participation without fair value distribution can simply create a new form of extraction.

The project therefore distinguishes: - paid work; - donated or sponsored contribution; - research participation; - in-kind materials; - shared infrastructure; - intellectual contribution; - future commercial rights.

The terms do not need to be identical in every project. They need to be explicit enough for participants to understand what they are contributing and what they may receive.

3.3 Stateless collaboration, persistent responsibility

The 28 July 2026 MCP specification introduced a stateless protocol core. In practical terms, each request can carry the information necessary to process it rather than relying on hidden transport-session state. If an application needs state across requests, it can use an explicit handle or identifier.

This is a powerful foundation for global engineering collaboration.

An engineer can reconnect from another client. A request can be routed through scalable infrastructure. Tool and resource catalogs can be discovered and cached. Workflows do not need to depend on a single fragile long-lived session.

But stateless transport does not remove real-world responsibility.

A physical system still needs persistent records for: - resource identity; - equipment state; - reservations; - work orders; - custody; - operation receipts; - authorization; - evidence; - payment; - exceptions and recovery.

The design principle is therefore:

Stateless request; persistent real-world responsibility.

This separation helps inclusion because participation is no longer tied to one hidden conversation or one client session. A new team can pick up a clearly identified work item, subject to authority, and understand what has already happened.

It also protects participants. A historical receipt should not be presented as a fresh observation. A retry should not create a second physical action. An expired permission should not be silently reused. A public resource description should not grant device-control rights.

This is how a scalable protocol can support accountability rather than weaken it.

3.4 Human Authority · Local Capability · Global Participation

HSIMC summarizes its participation model in three phrases.

Human Authority

People and organizations retain the right to define the permitted use of their resources, data, equipment and work.

Human Authority does not mean that every action requires manual clicking. It means that automation operates inside an explicitly delegated boundary and that responsibility can be traced.

Local Capability

Useful physical capability exists in specific places.

A machine, room, operator, material batch or delivery service has a location, availability window, condition and owner. Global AI cannot erase those facts. It can help coordinate them.

The local nature of physical capability is not a limitation. It is the source of much of its value.

Global Participation

Design, analysis, translation, software, simulation, quality review and business collaboration can come from anywhere, provided that contributors work against a common interface and evidence boundary.

Global participation does not require all data to be public. It requires a way to expose the right information to the right participant under the right authority.

Together, these principles support a network in which a small contributor can participate without pretending to be a global factory, and a global team can use local capability without pretending that physical execution is merely an API call.

3.5 The SDGs as a lens, not a marketing badge

The 17 Sustainable Development Goals provide a broad framework for considering social, economic and environmental outcomes. This white paper will map all 17 goals, but relevance will differ by work item.

The strongest early connections are expected to include: - SDG 4 - Quality Education: learning through real, reproducible engineering work; - SDG 8 - Decent Work and Economic Growth: paid participation, small-enterprise capability and productive work; - SDG 9 - Industry, Innovation and Infrastructure: enabling smaller participants to connect to modern engineering and production systems; - SDG 10 - Reduced Inequalities: reducing barriers created by geography, language, capital, age or access to infrastructure; - SDG 12 - Responsible Consumption and Production: repair, reuse, material evidence and lifecycle awareness; - SDG 17 - Partnerships for the Goals: practical multi-party cooperation around technology, knowledge and implementation.

A responsible SDG mapping requires a claim boundary.

For every proposed contribution, the project should ask:

  1. What changed compared with the baseline?
  2. Who benefited?
  3. Who carried additional cost or risk?
  4. What evidence supports the result?
  5. What remains unknown?

This prevents the SDGs from becoming decorative icons attached to ordinary business activity.

3.6 ESG as an operating discipline

ESG is useful in this project when it shapes decisions, not when it is added after the work is finished.

Environmental

The project should measure relevant material and energy changes, including: - new material introduced; - waste avoided or created; - repair and reuse; - process energy; - transport; - device and battery lifecycle; - end-of-life handling.

A recycled input is not automatically a lower-impact outcome. The full process matters.

Social

The project should observe: - who can participate; - who receives paid work; - whether new skills are learned; - whether physical burden or waiting is reduced; - whether access improves for people with different needs; - whether workers gain or lose agency; - whether the system shifts risk onto less powerful participants.

Governance

The project should preserve: - resource ownership or responsibility; - authority boundaries; - version history; - acceptance evidence; - payment and contribution records; - escalation and correction paths; - privacy and data boundaries.

The same evidence that helps engineering reliability can therefore help ESG credibility.

3.7 The smallest meaningful global action

A global ecosystem can sound impossibly large. The project deliberately defines a much smaller starting action.

A participant may bring: - one material sample; - one manufacturing process; - one machine hour; - one jig; - one camera image; - one software adapter; - one test method; - one local site; - one logistics window; - one real customer need.

The participant then describes: - what can be provided; - where and when; - who is responsible; - what use is permitted; - how the result can be checked.

That is enough to start.

The network becomes real when a second party can use the contribution, produce an accepted result and return evidence.

The next level is independent reproduction at another place.

The level after that is repeatable economic value.

This creates a ladder of participation that does not require every contributor to become a software company, own expensive infrastructure or understand the whole system before they can add value.

3.8 Part I conclusion - from inclusion to capability

The opportunity described in Part I is larger than hardware and smaller than ideology.

It is larger than hardware because the central problem is not a single card, computer, robot or protocol. It is the coordination of people, tools, materials, machines, authority and evidence around useful physical outcomes.

It is smaller than ideology because the proposal can be tested one function at a time.

Can a local process produce a sample that another team can verify?

Can a remote engineer work before final hardware ships?

Can an existing machine gain a useful new interface without being replaced?

Can a person with an ordinary tool contribute a measurable function?

Can the value, cost and responsibility be understood by everyone involved?

If the answers become yes, the idea gains substance.

Part II will move from the opportunity to the starting materials themselves: stone, thread, printing and other familiar processes, and the engineering discipline required to turn them from interesting possibilities into reproducible capability.


Evidence boundary for Part I

This is an independent HSIMC white-paper proposal. It does not claim endorsement by the United Nations, OpenAI, the Model Context Protocol project, ISO or any named ecosystem partner.

Current inherited engineering evidence remains unchanged: - C0-005: documentary T736 hardware baseline; - C0-007: source-verified simulator; - C0-008: historical Azure remote simulation record, 14/14 on 10 November 2026; - H0-001: no new physical T736 probe; the physical reference board remained unbound in that handoff; - the credit-card-sized hardware remains a design target rather than a finished-product claim.

Exploratory mobility, materials and local-manufacturing examples are used to define questions and collaboration patterns. They are not presented as validated product specifications unless a specific evidence level is stated.

Selected source notes

  1. OpenAI, Accelerating engineering for the physical world, 3 November 2026. https://openai.com/index/accelerating-engineering-for-the-physical-world/
  2. Model Context Protocol, The 2026-07-28 Specification. https://blog.modelcontextprotocol.io/posts/2026-07-28/
  3. Model Context Protocol, specification dated 28 July 2026. https://modelcontextprotocol.io/specification/2026-07-28/
  4. United Nations, Transforming our world: the 2030 Agenda for Sustainable Development. https://sdgs.un.org/2030agenda
  5. United Nations Sustainable Development Group, Leave No One Behind. https://unsdg.un.org/2030-agenda/universal-values/leave-no-one-behind
  6. ISO/IEC 7810, Identification cards - Physical characteristics; ID-1 nominal dimensions are 85.60 mm x 53.98 mm. https://www.iso.org/obp/ui/#iso:std:iso-iec:7810:ed-3:v1:en
  7. HSIMC, Card Phone and Cloud Development - Cross-Project Hand Off H0-001, 11 November 2026.
  8. HSIMC, White Paper Executive Summary W0-004, 11 November 2026.