If you work in education, you have probably heard some version of the same conversation this year.

“We really need an AI policy.”

Then comes the inevitable response.

“We are still waiting for the regulations.”

That second sentence is becoming harder to justify.

There has been a surprising amount of movement around AI and education across Europe during 2026. Some of it has been buried in regulatory documents, some of it has arrived through education policy, and some of it has happened at national level rather than in Brussels.

The important thing is that schools now have considerably more to work with than they did a year ago.

Here are five of the biggest myths I keep hearing, and what is actually happening.

“There is still no proper guidance on AI in education.”

There is now.

The Council of Europe introduced its Compass for AI and Education in Strasbourg in October 2025, providing a framework around AI literacy, teaching and learning with AI, evaluation, and regulation and governance.

Since then, the work has continued.

In September 2026, the Council of Europe adopted Recommendation CM/Rec(2026)12 on Artificial Intelligence Literacy.

The recommendation puts education at the centre of Europe's response to AI and looks beyond technical skills to questions around human rights, democracy, critical thinking and social agency.

That is a long way from saying that nobody has written anything yet.

“The EU AI Act's education rules arrived in August and schools now have to comply with everything.”

Not quite.

The timetable changed.

Following the EU's AI Omnibus, the rules for high risk AI systems covered by Annex III are now scheduled to apply from 2 December 2027.

That is an important delay, but it is not the same thing as the underlying requirements disappearing.

The AI Act still identifies certain educational uses as high risk, while other provisions are already applicable.

So, if your school interpreted the delay as “we can forget about this for another year”, it may be worth looking again.

“This is all Brussels policy that has very little to do with teachers.”

This year has made that argument harder too.

In May, the Council of the European Union approved conclusions on teachers in the era of artificial intelligence.

The message was fairly clear: teachers should not simply be treated as users of AI tools handed down by somebody else.

The Council called for teachers to have opportunities to contribute to the design and evaluation of AI tools, alongside stronger AI literacy, appropriate training and safeguards around workload and wellbeing.

Spain has been moving in a similar direction.

The Consejo Escolar del Estado's AI education framework was developed through collaborative work connected with the Gijón 2026 meeting and provides guidance for schools, educational communities and school councils.

Again, the direction is towards shared discussion rather than simply handing schools a list of technology rules.

“Spain is behind the rest of Europe.”

The reality is more complicated.

Spain established AESIA, the Agencia Española de Supervisión de Inteligencia Artificial, with its legal foundations created in 2023.

Spain is also working on its own national AI governance legislation.

So it would be misleading to say Spain has finished its AI regulation.

It would also be misleading to say that Spain has not engaged seriously with AI governance.

The country already has a dedicated supervisory agency, a national legislative process and education specific work happening alongside the European framework.

“AI policy and safeguarding policy are separate issues.”

If you work in an English school, this one deserves particular attention.

The 2026 edition of Keeping Children Safe in Education came into force on 1 September.

The Department for Education's digital safety standards now explicitly tell schools to consider AI generated content when reviewing their filtering and monitoring arrangements.

That matters because it makes the relationship between AI and safeguarding much more difficult to ignore.

An AI policy cannot simply sit in a digital strategy folder while safeguarding, online safety, behaviour and child protection policies carry on as though AI does not exist.

The issues increasingly overlap.

So what should schools actually do?

This is where the conversation becomes more useful.

Schools do not need to predict every regulatory change that might happen between now and 2027.

They do need to know what they currently allow, what they do not allow, what information staff and pupils can put into AI tools, how AI can be used in assessment, how staff are expected to disclose or acknowledge AI use, how new tools are approved, and how AI related safeguarding concerns are handled.

They also need to decide who owns the issue.

Is AI simply an IT matter?

Is it curriculum?

Is it assessment?

Is it safeguarding?

Is it professional development?

In reality, it is becoming all of these things.

The regulatory picture is still changing, but the excuse that “there is no guidance yet” is becoming increasingly difficult to sustain.

There is guidance.

There are frameworks.

There are deadlines.

Some deadlines have moved, some rules are already active and some national legislation is still being developed.

The sensible approach for schools is therefore not to wait for everything to become final.

It is to build a proportionate policy now and keep it capable of changing as the regulatory picture develops.

Because the question facing school leaders in 2026 is no longer:

“Has anyone written guidance on AI in education?”

It is:

“Have we actually read it, and do we know what it means for our school?”

If you are working through this at school level, we have put together a free AI Policy Checklist covering the areas schools should be thinking about now.

And if you are not sure where your school currently stands, we also offer a free 30 minute discovery call to help separate the genuinely urgent issues from the developments that can simply be monitored for now.