7. Early Identification of At-Risk Pupils
Description of the case
Who submitted this as a case?
The case was submitted in connection with the application for MedFin funding.
Brief description of the case (what is this / what does it concern?)
This case concerns identifying pupils who are at risk of «dropping out» as early as possible, so that measures can be put in place and so that measures and relevant information follow the pupils through their education and help keep them from dropping out.
It is thus about finding indicators of pupils at risk, choosing relevant measures, and ensuring that key data is shared with all actors in the ecosystem around the pupil.
What problem does the case address / What works poorly today?
“Everyone included – no one left outside” is a vision that shapes Norwegian policy for childhood, youth and schools. The aim is to ensure that all children and young people experience belonging, mastery and security, and to prevent exclusion caused by bullying, school absence or a lack of social connection.
The challenge receives a great deal of political attention, and work is being done both in research communities and across the sector to understand the challenges and develop measures. The most important measures, divided into four main areas, are:
- Early intervention and academic follow-up
- Early assessment: Identify reading, writing or numeracy difficulties as early as the first year of school.
- Rapid response: Put intensive support measures in place as soon as a child falls behind.
- Adapted education: Give schoolwork that matches the pupil’s level of mastery to avoid a sense of failure.
- Practical learning: Vary teaching with more practical and vocationally oriented methods for pupils who are tired of school.
- Psychosocial environment and well-being
- Belonging: Create a safe classroom environment where everyone feels included and seen.
- Bullying prevention: Have zero tolerance and concrete action plans against bullying and exclusion.
- Relationship building: Ensure that every pupil has at least one safe adult at school whom they trust.
- School health service: Have easily accessible public health nurses, psychologists and social workers at the school.
- Follow-up of absence
- Concrete absence routines: Contact with the home in the case of unauthorised absence.
- Worrying absence: Put measures in place as soon as absence exceeds 10–15 per cent.
- Mutual plans: Draw up “return plans” together with the pupil for those who have been away for a long time.
- Interdisciplinary collaboration and the home
- Close collaboration with parents: Maintain an ongoing, positive dialogue with parents and guardians, not only when something is wrong.
- The BTI model: Use the “Better Interdisciplinary Effort” (Bedre Tverrfaglig Innsats) model to coordinate services such as school, PPT (educational-psychological service), child welfare and health.
- Transitions: Create a safe framework around the transition from primary school to lower secondary school and on to upper secondary.
Other initiatives the case touches on
Several ongoing initiatives border on this case, including the work on a national absence register and a core group for vulnerable young people. A pilot must map these and clarify the interfaces, so that SAMT-BU does not duplicate work that is already under way.
Where do breakdowns in information flow or responsibility occur?
A recurring challenge throughout the pathway is a lack of access to data for those who are to assess and implement various measures. Both legal and technical barriers contribute to decisions being made on too thin a basis, and often too late.
Things also often «slip» in the transitions between kindergarten, primary school, lower secondary school and upper secondary, or in transitions between municipalities/county authorities when a pupil moves. Key information that is absolutely essential for the right measures to be put in place at the «new» place is not transferred, and as a result the pupil does not get the follow-up they deserve.
Who are the primary actors affected
- Pupils
- Parents and guardians
- Staff at educational institutions
- Staff in support services (state and municipal)
What are the consequences for the actors of the current situation.
Pupils do not get the support they are entitled to and need in order to complete their education. In the worst case, this can mean that the pupil «drops out» and ends up living a life of exclusion.
What could a desired situation (dream journey) look like?
The ideal is of course that information flows freely between the actors in the ecosystem around the pupil, but in such a way that privacy is protected and the information does not help entrench a bias. Knowing that someone is «at risk» can in itself help make it a self-fulfilling prophecy.
The following insight work has been carried out
SAMT-BU does not have a full overview of the insight work that has been carried out on this issue, but we are aware that a number of studies, research projects and user studies have been carried out on exclusion and what it takes to prevent pupils from «dropping out». An important task for a pilot will be to identify what is relevant for us.
Which of the project’s objectives do we consider the case to address?
This case touches the very core of the issue that was the starting point for the project: How can better use of data across organisations contribute to better services for pupils and parents?
The issue this case touches on is, however, very broad; a great many people are working on similar assignments, and there is at times considerable political disagreement about some of the measures. It is therefore important that SAMT-BU scopes the case to concern things we can actually do something about, and where we do not get in the way of many others.
Taking data as a resource as the starting point, there are mainly two areas where SAMT-BU can contribute:
- Early identification of at-risk pupils
- Sharing of data between the actors, especially in connection with transitions in the educational pathway
By working on these, we will be able to contribute to the objectives:
- Shared information models and standardised interfaces that over time cover «all» areas and sectors.
- Reusable solutions and shared solutions
- Simpler user journeys for staff, parents and guardians, pupils and students throughout the learning pathway - from kindergarten to primary and lower secondary school, upper secondary and higher education.
- Collaboration and innovation across public and private actors in the ecosystem, and increased competition in the market.
- Long-term goals of lifelong learning, etc.
- Scaling and transfer of the project’s learning and outputs to other areas and sectors, following piloting in this project.
A little about implementation
Early identification of at-risk pupils
There is reason to believe that it should be possible to develop an AI-based prediction model that can identify, with a high degree of certainty and at an early stage, pupils who are at risk of «dropping out». The broader the data foundation the model is built on, the more accurate it can become. Such a model would of course include data from kindergarten and school on absence, test results, etc. In addition, one could consider including demographic and socio-economic data and other data about the family. It will also be necessary to obtain historical data to train the model on.
Besides the purely legal challenges, such a model is naturally not free of ethical problems, and one must avoid creating a «black box» AI that cannot explain why its conclusions are what they are. Strict rules must also be drawn up for how the model can be used, and control mechanisms must be developed so that it is always a human who ultimately decides whether measures are to be put in place. However, the ethical dilemmas will primarily relate to how a possible model is applied, and not to the model itself. In SAMT-BU we can therefore focus on the latter.
A pilot run by SAMT-BU will therefore initially be of an exploratory nature and must find answers to:
- Is it possible to create a prediction model?
- What data is needed for it to be accurate enough?
- What technical and legal obstacles may need to be removed for the model to be usable in practice?
Law is central to this case, and the question is not only which obstacles must be removed. It is just as important to examine how the current regulations can be used – what is actually possible to achieve within the current framework? Experience shows that the room for manoeuvre is greater than the widespread uncertainty in the services would suggest.
In carrying out the work, we should consider collaborating with KI Norge and AI Learn and/or AI Decisions, or possibly other research communities within the local government sector.
Sharing of data between the actors, especially in connection with transitions in the educational pathway
Several of the other registered cases in SAMT-BU touch on this issue (see https://docs.samt-bu.no/en/behov/use-cases/):
- CASE 2 – Seamless transition from primary and lower secondary school to upper secondary education, which was carried out as Pilot 1 in the period 01.02.2026 - 30.04.2026. We are now considering continuing work on it from 01.09.2026.
- CASE 4 – Relocation of pupils between municipalities/county authorities during the school year, which was started just before the summer and will run until October 2026.
- CASE 10 – Transition to other municipal services, which we are considering starting from 01.09.2026. (This also covers Case 18, which is based on input from the Education Agency (Utdanningsetaten) in Oslo municipality)
- CASE 29 – Transition from kindergarten to primary school, which has not yet been assessed
- CASE 30 – Transition from primary school to lower secondary school, which has not yet been assessed
We believe that once several of these have been carried out, there will be a foundation that can be built on further, including other relevant data sources from actors such as STAT-PED and NUBU. For the time being, we therefore do not wish to work on this part of the case.