AI for school counselors: uses, limits and student data

AI is mostly useful for preparation: it gives you a first draft that you review, correct and make your own.

Some uses that genuinely save time: drafting the script for a tutorial session on decision-making; adapting one text about study routes for a 13-year-old, a 15-year-old and a family who doesn’t know the system; preparing a term’s tutorial plan from your goals; summarising a long regulation into the points that affect your school; or brainstorming questions for a guidance interview with a student who doesn’t know what they want.

In all of these you’re working with ideas and general material, not with data about a specific student. And in all of them the last word is yours: AI makes mistakes, sometimes invents things and doesn’t know your school.

AI shouldn’t decide anything about a specific student: not which route suits them, not what their guidance report says, not whether they need support.

A tool can organise information or suggest options, but it doesn’t know the student’s story, what they said with a look in the last interview or what’s happening at home. Decisions about their path need someone who knows them and answers for those decisions.

The European rules also treat exactly these uses with special care. When AI influences access to studies or how a student is assessed, it is classed as high-risk and requires strict controls, starting with human oversight.

With a public chatbot, no. Don’t paste data that could identify a student, and certainly not health data, special educational needs or their family situation.

What you type into a public tool may be stored outside your school and used for purposes nobody has authorised. With minors, and with data this sensitive, the risk isn’t worth it.

If you need help with a case, remove everything that identifies the student and keep the general situation: “a Year 10 student with good science grades who is torn between academic upper secondary and a vocational course”. To work with real data, use only tools your school or education authority has a contract for that covers data protection.

When in doubt, ask your data protection officer. Every school has one, either its own or assigned by its education authority.

The European AI rules class four uses in education as high-risk, and their obligations will apply from December 2027.

Those four uses are AI that decides access or admission to studies, AI that evaluates learning outcomes, AI that assesses what level of education a student can reach, and AI that monitors students during exams. For these, the rules require, among other things, human oversight, transparency and control over data quality.

The original date was August 2026 and it has been pushed back to December 2027, although many Spanish-language pages still quote the old one. The delay doesn’t change the substance: if a tool at your school does any of those four things, it’s worth asking now how it complies.

Yes. Since February 2025, the European AI rules ban using AI to recognise people’s emotions in schools, except for medical or safety reasons.

That includes tools that claim to read from a face, a voice or the way someone writes whether a student is paying attention, sad or nervous. In guidance, you learn about emotions by talking with the student, not by measuring them.

It’s a good idea. Since February 2025, the European rules have asked that people who use AI at work have enough knowledge to use it well, and in 2026 that duty was softened towards supporting and enabling that training.

In practice, the essentials are enough: knowing that AI makes mistakes and sometimes invents things, knowing which data not to share, always reviewing what it produces and understanding what each tool at your school does. Ask your school what training it offers.

The conversation, the family context, judgement and the decisions about each student’s path are still yours.

AI can give you back time for what matters most: sitting down with a student, hearing what doesn’t show in their grades, talking with their family and supporting a decision that is theirs. That is guidance, and no tool does it for you.

That’s how we think about it at UTalento: student data stays inside the school, and the guidance team reads and interprets it themselves, so they arrive at each conversation better prepared.