Documentation

The documentation behind bringing AI into the classroom with judgment.

This page summarizes the operational material available as of 13 April 2026 for Lira AI pilots. The full documentation and its institutional versions are shared on request during the validation process.

The point is not to publish hot air. It is to show that the pilot already has scope, safeguards and method before it scales.

Pilot baseline document

Defines scope, participants, authorizations, suspension and the conditions for working with students.

Go to section

Classroom use framework

Sets out how AI is integrated to guide processes without replacing thinking or teacher supervision.

Go to section

Ethical guidelines

Covers human oversight, protection of minors, permitted use and institutional limits.

Go to section

Evaluation methodology

Explains what is measured during the pilot and what evidence unlocks an institutional phase.

Go to section

Section

Pilot baseline document

The operational documentation already describes a controlled, bounded, supervised and reversible pilot. The idea is not to open access indiscriminately, but to agree with each institution on what is tested, with whom, and within what limits.

Defines scope, classes, subjects, dates and which features are enabled.
Sets clear roles between the institution, Lira AI and the responsible teachers.
Allows early suspension or termination for pedagogical, legal, technical or privacy reasons.
Requires institutional sign-off before any activity involving students.

Section

Classroom use framework

Lira is not positioned as an answer generator. The classroom approach starts from a simple premise: AI is worth having when it helps students think better and makes the process visible, not when it replaces the student's work.

Activities are designed to surface intermediate steps and reasoning.
AI is built into the assignment, not a shortcut around it.
The teacher keeps the context, the criteria and the ability to step in.
Traceability shows where a student got stuck, what they tried, and how they progressed.

Section

Ethical guidelines

The institutional documentation sets limits from the start. The pilot does not authorize automated decisions with significant impact, nor does it hand professional teaching judgment over to the tool.

Every AI output remains subject to reasonable human oversight.
No work with minors begins without the corresponding institutional review.
The scope of the pilot cannot be widened unilaterally.
Pilot data is not reused for purposes incompatible with the authorized implementation.

Section

Evaluation methodology

The pilot is evaluated against observable evidence. The current methodology combines pedagogical and operational signals to answer whether Lira works in a real classroom and whether it can support a next phase.

Student-AI interaction and support of the thinking process.
Real usefulness for the teacher and how smoothly it fits into class.
Adoption, activation, operational quality and pilot efficiency.
Aggregate findings, not isolated impressions, before deciding to scale.

Institutional access

Request the full documentation package

If your institution wants to review the pilot in detail, we can share the operational documentation, privacy terms, methodology and implementation scope.