Anyone who works with minors says they care about protecting them. A statement cannot be verified; an architectural choice can. This page lists the technical decisions we have made, so that they can be checked.
Students answer and interact only through structured choices. It is not a rule of behaviour that someone could get around: it is the shape of the interface.
It is an architectural decision that structurally reduces the risk of personal data or inappropriate content being entered while interacting with the tutor, and keeps the session within the learning perimeter.
The lesson and the questions are drawn exclusively from the pages, handouts or notes the student uploads. Never from generic content, and never from an archive of ready-made topics.
Adaptation is based on objective learning data, such as the age group, the subject and the characteristics of the uploaded material, and on the preferences the student chooses explicitly.
The type of specific learning disorder is not needed to use the platform and is never used to adapt lessons: families may state it, if they wish, only so that the student can save the text extracted from the uploaded pages, as copyright law allows. We do not use behavioural signals observed during the session to assign profiles or to change the level automatically: the path follows understanding, not a categorisation of the student.
After a mistake it rephrases and presents the concept again, guiding the student to another attempt. A system that solves problems for the student produces finished homework and no learning.
Text size, spacing, colour contrast and speech synthesis can be changed during the session, according to the reading preferences of the person studying. It is not a special mode to switch on: it is how the platform normally works.
“If the answer is wrong, the responsibility lies with the platform, not the student.”
It is the tutor that rephrases. Not the student who has to justify themselves.
The systems receive the learning content needed for processing, not the name of the person studying.
Application, database, backups, error monitoring and speech synthesis run in European regions. Backups are encrypted, and the restore procedure is tested regularly.
Processing that involves providers based outside the Union is governed by the safeguards required by the GDPR, with data processing agreements signed case by case.
The DPO can be contacted directly at dpo@menthorai.ai, to exercise the rights provided by the GDPR and for any report.
MENTHORAI does not host advertising, does not share data with third parties for commercial purposes and has no interest in keeping students on the platform any longer than necessary.
Data protection by design and by default: minimisation of the information collected, pseudonymisation, particular attention to the protection of minors. A data protection impact assessment accompanies the development of the platform.
The Italian ministerial guidelines on specific learning disorders list concept maps among compensatory tools, and specify that these tools relieve students of a task made difficult by the disorder, without making the task cognitively easier. Placing the concept map at the end of the path, and not in place of it, is our methodological choice in that direction.
Students are clearly informed, in words suited to their age, that they are interacting with an artificial intelligence system, as required by Article 50.
The platform is designed in line with the Italian Ministry's guidelines for introducing artificial intelligence in schools, and supports everyday study consistently with the student's personalised learning plan where there is one. The guidelines ask that artificial intelligence support the acquisition of authentic skills without replacing students' effort, reflection and autonomy: it is the principle on which the tutor is built, and the reason it never gives a ready-made answer.
The articles of association state that the company's activities are not healthcare activities and do not involve direct medical services. MENTHORAI is a learning tool.
Transparency works both ways. We state just as clearly what this tool does not do.
Protocol 6 and the BLIP method are proprietary methodologies deposited with SIAE, the Italian Society of Authors and Publishers. Registration is not meant to lock knowledge away: it allows us to state publicly which method the platform is based on, and makes its authorship verifiable.
MENTHORAI works with schools, universities, clinical centres and associations to validate and disseminate the technologies it develops. Requests for methodological information are welcome: info@menthorai.ai.
MENTHORAI