Trust, Privacy and Accessibility

GLEOS is designed to make education, recruitment and Japan pathway support easier to understand while respecting privacy, accessibility, security and informed decision-making.

Accessibility

Designed for different users

GLEOS should be usable by people with different devices, abilities, languages and levels of digital experience. The platform should remain understandable whether someone is using a desktop computer, a mobile phone, a keyboard, a screen reader or a slower internet connection.

Clear and usable interfaces

Pages should use readable text, clear headings, sufficient colour contrast and meaningful labels. Important actions must not depend only on colour, hover effects, animation or small icons.

Forms should explain what information is required, show errors beside the relevant field and preserve entered information when a correction is needed.

Keyboard and assistive technology support

Users should be able to navigate menus, forms, tables, drawers and dialogs with a keyboard. Focus should remain visible and should return to the correct control after a drawer or dialog closes.

Important changes, such as a saved result or an updated application status, should be announced in a way that can be understood by screen-reader users.

Responsive and multilingual access

The website and portal should work on small screens from 320 pixels wide without sideways scrolling. The interface should support reduced-motion preferences and provide text alternatives for meaningful images and media.

GLEOS supports English, Bengali and Japanese where translations are available. Language changes should not erase form data, filters or the user's current position.

Reporting an accessibility barrier

If a user cannot complete an action or understand a page because of an accessibility barrier, they should be able to contact GLEOS for assistance.

Reported barriers should be reviewed as part of future platform improvements.

AI boundary

What AI may do

AI may help explain information, organise approved GLEOS data, provide preliminary eligibility guidance and suggest practical next steps. It may also help present management information in a clearer format.

What AI cannot decide

AI does not make final decisions about admission, pathway approval, course completion, job selection, visa approval, COE approval, departure or candidate rejection.

An AI result is not a visa decision, legal opinion, job guarantee or immigration guarantee. Users must always have access to the normal GLEOS process and human counselor support.

Reliable AI behaviour

AI responses should be based on approved GLEOS information. The system should not invent requirements, costs, employers, vacancies, processing times or guarantees.

When the AI service is unavailable or produces an unreliable response, GLEOS should use a safe rule-based explanation instead. Where appropriate, authorised staff should be able to review AI-assisted guidance.

Security boundary

Server-enforced access

GLEOS protects information through server-side permissions. A visible button or page link does not determine whether an action is allowed. Every sensitive action must be checked again by the server.

Users should only access information required for their role. Candidates should see their own information, counselors should see information within their assigned scope and staff should not access private details without permission.

Account and data protection

Passwords, tokens, provider credentials and API keys must be protected and must never appear in browser code, public pages or ordinary logs.

Documents should use secure access links and existing malware-scanning controls. Sensitive endpoints should use authentication, rate limits and audit records.

Monitoring and recovery

Important actions such as approvals, corrections, status changes, document reviews and financial activity should be recorded in the audit history.

Major releases should use verified backups and rollback procedures. Security problems should be investigated without exposing personal information unnecessarily.

Evidence boundary

Separating fact from guidance

GLEOS distinguishes between information supplied by a candidate, information verified by staff, results recorded by teachers, documents uploaded by candidates, documents reviewed by staff, employer-provided information and AI-generated guidance.

The platform should make it clear when information is missing, self-reported, awaiting review or officially verified.

Verified records

A candidate's self-declared Japanese level is not automatically treated as a verified JLPT result. A submitted document is not automatically an approved document. A job listing does not guarantee selection, employment or departure.

Training progress should come from teacher-recorded classes, lessons, assignments, tests and outcomes rather than from a student's own claim.

Traceable decisions

Important decisions should remain traceable through the responsible staff member, date and time, original status, updated status and reason for any correction.

This allows candidates, staff and management to understand what happened, which evidence was used and why the current status exists.

No unsupported promises

When evidence is incomplete or a decision depends on further review, GLEOS should say so directly. The platform should provide the next practical step instead of presenting assumptions as confirmed facts.