Security, governance and cognitive agency

Cross-cutting principles

Protection is necessary. Positive agency matters too.

Governance should prevent misuse while also giving workers, learners, service users, clients, researchers and other participants meaningful ways to see, influence and contest the system.

Purpose limitation

Collect and use data for the stated support, measurement, research or pilot purpose—not unrelated profiling, selection or disciplinary monitoring.

Data minimisation

Prefer the least data needed for routing, delivery, progress, evaluation and support. Keep commerce, identity and cognitive performance logically separate where practical.

Participant authority

Explain what is measured, what is inferred, what remains uncertain, who may see it and what the participant may decline, correct, export or request to be deleted.

Contestability and review

Provide a route to correct inaccurate evidence, challenge a material inference and request review by an accountable person before any consequential use.

Worker- or learner-controlled evidence

Personal development records and evidence of mastery should be private by default and shared deliberately for a defined purpose, rather than becoming an unrestricted institutional profile.

No consequential automation

Do not use app scores, engagement, Zone outputs or provisional CSI inferences for automated employment, admissions, progression, clinical, insurance or eligibility decisions.

Access boundaries

Separate participant, coach, educator, practitioner, researcher and institutional roles; expose only what each authorised role needs for the stated purpose.

Version traceability

Record the protocol, content, model, route-contract and scoring versions used when a result, recommendation or interpretation was produced.

Voice and procedural agency

A concern should enter an accountable route—not disappear into a survey.

A future CSI participant surface should let a person raise a pressure point, choose an appropriate visibility level, see where it was routed and receive a reasoned response or review path.

01Record the observation or pressure pointEvidence
02Choose private, named, aggregated or protected visibilityControl
03Route it to the appropriate owner or formal governance channelAccountability
04Receive acknowledgement, action, explanation or a reasoned refusalResponse
05Correct, contest, appeal or request further human reviewReview

CSI can structure evidence and routing. It does not replace formal grievance, whistleblowing, complaints, appeals, safeguarding, data-protection, professional-governance or representative processes.

Sector-specific routes

The same principle requires different governance channels.

Sector is context. A concern about workload, AI, learning support, service quality or safeguarding must be routed through the authority structure appropriate to the setting.

Setting	Illustrative routes	Protected boundary
Employment and professional work	Process owner, line management, people development, health and safety, occupational health, AI/data governance, grievance, protected disclosure or employee representation.	No app score or CSI inference becomes a promotion, disciplinary, redundancy or suitability decision.
Education and research	Tutor or programme lead, learning support, assessment appeal, research supervision, safeguarding, data protection, academic governance or student representation.	No cognitive-app score becomes an admissions, grading, progression or disciplinary score.
Health, care and behavioural support	Practitioner, service lead, complaints route, safeguarding, clinical or professional governance, consent review or data protection.	CSI remains decision support and does not replace clinical, care or safeguarding judgement.
Sport, coaching and performance	Coach, performance lead, participant advocate, safeguarding, medical or professional governance and programme review.	Participation, selection and return-to-performance decisions remain governed human decisions using appropriate evidence.

Website route builder

Process locally first.

The current deterministic route builder operates client-side and shows the result without requiring contact details.

Analytics boundary

Aggregate funnel events may be measured
Do not send detailed workflow or learning answers to ad platforms
Do not send health-context answers, cognitive profiles or voice submissions to ad platforms
Do not use optional AI-context tags for advertising segmentation
Do not infer non-participation as low motivation or low capability

Before an institutional pilot

Questions that must be answered.

Area	Required decision
Controller and processor roles	Name the legal entities and responsibilities for each data flow.
Lawful basis and transparency	Document why data are processed, what participants are told and what is optional.
Participant view and correction	Define what participants can see, correct, contest, export, delete or ask a human to review.
Voice and representation	Define how pressure points can be raised, aggregated, routed and escalated, including representative routes where relevant.
Development and progression boundary	Keep private development support and participant-controlled mastery evidence separate from selection or progression decisions.
Retention and deletion	Set periods by data type and a deletion process.
Hosting and subprocessors	List regions, providers, transfers and contractual safeguards.
Access and reporting	Define who can see individual versus aggregate data.
AI delegation and provenance	State what AI may do, what people must still understand, what requires verification and who retains final authority.
Incident response	Specify reporting, containment, correction and notification responsibilities.
Decision boundary	Prohibit automated or unsupported employment, educational, clinical, service-eligibility or performance decisions based on provisional outputs.

Need a pilot governance conversation?

Define the participant rights, authority, reporting and contestability model before representing an institutional route as production-ready.

Discuss governance