Frontier case 03 — Neurotechnology / Identity

Human–machine interfaces

If interfaces between people and machines become more capable, what architecture keeps identity, autonomy and access from depending on a single provider?

Draft v0.1 — no prior versions Updated 2026-09-23
Frontier vision

Brain–computer interfaces may increasingly connect people and machines. Current clinically demonstrated use is largely assistive and narrow; broader capability is proposed rather than established.

Assumption under examination

That the identity, consent and data layer of such a system can be architected to survive independently of whichever company supplies the device.

Ours, not theirs. This is the thing being tested.

Research question

If human–machine interfaces become more capable, what architecture protects identity, autonomy, privacy and access?

Status

Draft

Frontier builders ask

FUTURE 2100 asks

Can humans connect directly with machines?

What protects identity, autonomy and access if that becomes possible?

Why it matters

What the answer would change.

Every other layer this project studies — payments, records, coordination — can in principle be migrated between providers. If a layer reaches into cognition, the cost of provider capture is different in kind, not degree. The architecture question is worth asking before the capability arrives, because afterwards the answer is inherited rather than chosen.

What it depends on

The variables, not a conclusion.

Identity
Whether the person’s identity is a record the provider holds, or one that exists independently of it.
Neural data
What is recorded, where it is stored, and whether it can be exported or deleted.
Consent
Whether consent is a single event or a revocable, granular, standing relationship.
Autonomy
What the person can refuse without losing the capability entirely.
Security
Failure modes when the layer is cognitive rather than financial.
Access
Who can obtain the capability at all, and at what price.
Ownership
Who owns the derived model, as distinct from the raw data.
Governance
Who decides, and by what process, when the rules change.
Provider dependency
What the person loses if the supplier fails, is sold, or changes terms.
Sub-questions

What we are actually asking.

  • Should the identity layer of a cognitive interface depend on a single company, platform or provider?
  • What would a portable identity layer for such a system even look like?
  • Can neural data be made exportable in a meaningful form, or is it inherently model-bound?
  • What happens to consent given years earlier, under different terms, by a person whose condition has changed?
  • What is the failure mode when the provider ceases to exist?
Affected systems

Where an answer would land.

  • Identity infrastructure
  • Medical device regulation and clinical governance
  • Data protection and cross-border data regimes
  • Insurance and employment screening
  • Assistive technology access and pricing
  • Digital rights frameworks
Alternative architectures

Unranked, on purpose.

Provider-held identity
The supplier holds identity, data and model. Simplest to build; highest capture cost if it fails.
Portable identity layer
Identity exists independently and the device authenticates against it. Harder; preserves exit.
Institutional custody
A clinical or public institution holds the layer. Shifts the dependency rather than removing it.
Self-custody
The person holds keys and data. Maximum autonomy; real usability and recovery problems.
Open specification with multiple providers
Interoperability by standard. Depends on a standard existing before the market consolidates.
Evidence

Each claim labelled by what kind of claim it is.

  • FactBrain–computer interfaces have been used clinically in assistive contexts, including restoring communication and control for people with severe motor impairment. This is established and narrow.
  • TrendInvestment and the number of active programmes in the field have grown. Growth in funding is not evidence of capability.
  • HypothesisThat an identity layer can be architected independently of the device supplier. Untested; this is the assumption, not a finding.
Counterarguments

The strongest case against us.

  • The architecture question may be premature by decades, and premature standards can entrench worse designs than none.
  • Neural data may not be portable in any meaningful sense, because it is only interpretable through the model trained on it. If so, portability is the wrong goal and the research should say so.
  • Clinical governance may already provide stronger protection than a technical architecture would, making this a regulatory question wearing an engineering costume.
  • Insisting on self-custody could restrict access for exactly the people who most need the assistive capability.
Unknowns

What we do not know.

  • Whether neural data is portable in any useful form.
  • Whether a standard could emerge before the market consolidates around one supplier.
  • What recovery looks like when the credential is tied to a person’s cognition.
  • Whether the access question and the autonomy question have compatible answers.
Experiments

Method first. Falsifier written before any result.

Port the question, not the data

Apply the existing provider-independence scale to a hypothetical cognitive-interface stack: score identity, data custody and derived model against the same L0–L5 criteria already used elsewhere in this project, and record where the scale breaks down.

FalsifierIf the existing independence model cannot express what is at stake for a cognitive layer, the model is insufficient and must be extended rather than reused.

The important one

What evidence would change our view?

A position no evidence could move is not research. These are the findings that would change this record — written before we have them, so they cannot be chosen afterwards to fit.

  • A demonstration that neural data is meaningfully portable between systems would strengthen the assumption considerably.
  • Evidence that portability is impossible in principle would reject the assumption as stated and redirect the work toward governance instead of architecture.
  • Evidence that independence requirements measurably restrict access to assistive users would force this project to state the trade-off openly rather than assume independence is costless.
Sources

Named, and marked unverified until a person has checked one.

Clinical literature
Peer-reviewed reports of brain–computer interface use in assistive clinical contexts
To be identified and cited individually · Source verification required unverified
Field survey
Published surveys of active neurotechnology programmes and funding
To be identified · Source verification required unverified
Related

Where this connects.