Digital twin for trainers: what it does today and what comes next

It is a version of the trainer made with artificial intelligence: it has their face and their voice, and a brain of its own holding their materials, their syllabus and their way of explaining things. Its job is to look after learners and prospective learners in the part of the work that repeats itself, always saying that it is an AI.

The short definition of Digital Twin by Legathum is this: a professional cloned with AI who serves clients and sells at any hour. For someone who trains professionals or company teams, that means questions about the syllabus, the doubts that arrive at eleven at night before an assignment is due, and the fifth explanation of the same concept no longer have to depend on you being available.

The part that repeats itself and does not require your judgement about a particular person. The table goes through the stages of a course and separates what the twin can handle from what stays with you.

The rule worth writing into your Book of Criteria comes from the last column: when a question asks for an assessment of a specific learner's work, the right answer is to refer it to you.

You do, if you created them, and that has consequences. In Spain, Article 10 of the Intellectual Property Act lists lectures, writings and teaching explanations among protected works, and Article 17 reserves to the author the rights of reproduction, distribution, public communication and transformation of their work. Other countries have equivalent rules.

For a trainer, this comes down to two practical rules. First: only materials you hold the rights to should go into your twin's brain, meaning your own or those you have permission to use in this way. A book by another author, slides from a school you once worked for or notes from someone else's course do not go in without permission. Second: if you train on behalf of a school or a company, check your contract first, because you may have assigned some of those rights.

In Digital Twin by Legathum, the brain belongs to you, and the Book of Criteria can be exported at any time.

Make decisions about them. Annex III of the EU AI Act treats as high risk AI systems used in education and vocational training institutions to decide access or admission, to evaluate learning outcomes, to decide the training a person will receive or to monitor students during exams. Under the timetable in Regulation (EU) 2026/1744, those obligations apply from 2 December 2027.

Outside an institution, the underlying idea still holds: a twin can explain, revise and guide with your materials, but the mark, the pass and admission to a course are your decisions. Nor should it monitor anyone or draw conclusions about how a learner feels.

And, as with anyone else, the learner has to know they are talking to an AI. Article 50 of the same Regulation requires this at the latest at the first interaction, and the twin says so in its first turn, with a name that ends in "Clon IA".

Very little. The GDPR requires data to be adequate, relevant and limited to what is necessary for the purpose. To answer a question about the syllabus, the question is enough; to prepare a session, what the course needs to know. There is no need to ask for personal details that do not help teaching, and learners should know what is kept from their conversations and why.

If you train company teams, add one precaution: what a learner asks the twin should not turn into a report for their manager without the learner knowing.

With what you already use to teach, put in order. Start with the syllabus and the learner guide, then add the glossary for your subject, the questions you are asked most and the exercises with their criteria. The clearer each concept is in your materials, the better the twin will explain it.

Then write your boundary rules into the Book of Criteria. For example: if a learner asks for the answer to an exercise that will be assessed, the twin guides them with questions and does not give the answer. Or: if a question falls outside the syllabus, the twin says so and suggests raising it with you in the next session. Each rule with its condition, its action, its reason and its source.

Nothing that has not been checked is presented as available. This table separates the two for a trainer.

Today consultancy is the profession the method and session types are designed for. A trainer who works with companies is close to that profile: explaining a method and supporting its application.

• Today: a twin with your face and your voice, a Master Brain with a Book of Criteria, the AI notice and the "Clon IA" name, verified and revocable consent, and a booking page.
• In development: training-specific workflows, with revision sessions, exercises and follow-up for each learner.
• In development: the twin delivering a paid service on its own and applying your rules while citing the rule behind each answer.