“Participation changes the question and improves the answer — when learning matters more than completion: Learning and technology”

Meaningful participation by students and families is not a final-stage reaction to a finished educational product. It is part of the way a school makes decisions. People who will live with a learning platform should be able to help define the need, test possible responses, examine harms, and see how evidence changes the result. This matters most when technology promises to sort information, recommend pathways, or automate contact. Educational value does not follow from a long feature list. It depends on whether the system supports learning, trust, and fair access to opportunity.

When those voices are present early, they expose questions that a procurement document can easily miss: who lacks reliable connectivity, which wording is inaccessible, what alert could create shame, or which digital process shifts administrative work onto an already stretched household. Participation changes the original question and can therefore change the answer. It does not transfer professional responsibility to children or families. It requires the institution to take seriously the situated knowledge they contribute.

Start with a shared account of the problem#

Technology purchasing often begins with an efficiency claim: fewer emails, quicker progress reports, or more complete records. None of those claims establishes an educational need on its own. Before looking for a vendor, a school can convene students, families, teachers, inclusion staff, counsellors, and data-protection specialists. Their first task is not to select an app. It is to describe, with concrete examples, what prevents learning or participation. The group should distinguish an access barrier from an information gap, an unfair decision, and a problem that may not require technology at all.

Technology purchasing often begins with an efficiency claim: fewer emails, quicker progress reports, or more complete records.

Participation must be possible for the people invited. Evening and daytime sessions, interpreters, short plain-language materials, reimbursement where appropriate, and non-digital ways to contribute prevent consultation from favouring those with the most free time. Article 12 of the UN Convention on the Rights of the Child recognises children’s right to express views in matters affecting them. In practice, a school must explain the decision in understandable terms and later show what happened to the views it received.

Scenario one: a dashboard that mistook activity for learning#

A secondary school considered a dashboard that grouped students by login time, submissions, and clicks. The vendor demonstration looked persuasive until a trial session with students revealed a different pattern. Some downloaded work at school to study offline; others completed tasks on paper before uploading them. Families explained that a single phone was shared at home and that the hour of connection said little about effort or difficulty. The dashboard was measuring a narrow trace of access, not engagement.

The question changed from how to identify disengagement to how to notice barriers worth discussing. The pilot removed individual rankings, added a simple way to record offline work, and prohibited use of the dashboard for discipline. During evaluation, each alert was compared with the student’s and tutor’s explanation. False positives were common enough that the central indicator was withdrawn. That outcome was not a failure of consultation. It prevented weak data from becoming a school label with real consequences.

Co-design needs decisions that can genuinely move#

Co-design becomes symbolic when the significant choices are already closed. Participants need influence over specific matters: what data is necessary, which roles can see it, how a recommendation is explained, how long data is kept, and when a feature must be disabled. The institution should state what it can change, what law or policy constrains, and who holds final accountability. A declared limitation is more trustworthy than an invitation that implies influence where none exists.

A reference group is not automatically the whole community. Membership should rotate, its selection process should be visible, and recruitment should seek experiences commonly absent from school technology decisions: disabled students, young people using more than one language, households with low connectivity, carers, and people with difficult prior contact with the institution. Accessibility testing is not an end-stage adjustment. It helps determine which uses are reasonable and which burdens should never be passed to learners or households.

Scenario two: booking support that excluded the people who needed it#

A school introduced an app for booking meetings and support sessions. Its first version required a personal email address, two-factor authentication, and a single morning slot. In a facilitated trial, parents and carers described shared phones and working hours that made the slot unusable. Students needing reading support found that instructions disappeared before they could finish them. The system was not merely inconvenient: it treated a particular pattern of access as a condition for speaking to the school.

The redesign added telephone and in-person booking, multilingual messages, longer reading time, and a way to authorise a carer without sharing credentials. It also removed the requirement to state the reason for a meeting before a slot could be obtained. Evaluation did not stop at the number of bookings. It compared drop-off rates, response times across channels, who gained appointments, and reported difficulties. The community had defined success as being able to ask for help without disclosing unnecessary information or clearing artificial hurdles.

Evaluate consequences, not only uptake#

A tool can show high use and still allocate opportunities unfairly. Participatory evaluation combines measures with accounts of experience: errors, waiting times, requests for help, clarity of explanations, ability to correct data, and episodes of embarrassment or exclusion. Results should be returned in accessible language and discussed in a setting where changes can be proposed. A percentage without context is not accountability, because affected people cannot test what it means or challenge the decision built on it.

Review thresholds should be agreed before deployment. If households using a particular language abandon a process more often, if one group receives more alerts without a pedagogical reason, or if a human alternative responds too slowly, the feature should pause. The EU Artificial Intelligence Act requires risk management, appropriate data governance, and human oversight for high-risk systems. In education, oversight becomes more than a formal role when it includes evidence from the people experiencing the system rather than only an internal vendor assessment.

Scenario three: recommendations that narrowed futures#

A platform suggested elective subjects from previous grades and logged activity. In evaluation workshops, students and families noticed that less demanding options appeared more frequently for young people who had recently changed language of instruction or had been absent for a long period. The system treated a temporary circumstance as a stable limit. Average accuracy did not answer the central question: what future did that recommendation open or close for this particular student?

The school suspended automatic recommendations while it reviewed variables and documentation. In the revised version, every suggestion displayed alternatives, the factors considered, and a route to correct data; any consequential choice required a conversation with guidance staff. Differences by access to support, language, and mobility were examined with privacy safeguards. The tool became preparation for a conversation rather than a predetermined route. Useful explainability is not a technical definition alone; it is a practical ability to question and revise what the system produces.

Scenario four: a notification that turned delay into blame#

At a primary school, an app automatically sent an alert when an assignment was not recorded as submitted. Families reported that messages arrived before a child could explain an absence or connectivity problem, and that the wording suggested a lack of care. Students described having to account for themselves in front of classmates the next day. Review showed that notification speed was producing more tension than capacity to respond.

The school replaced the automatic warning with a discreet delayed reminder that a teacher could review, while retaining human contact for situations needing support. Families helped draft the wording and test its frequency. This apparently small intervention changed the evaluation standard. It was no longer enough to show that messages were delivered. The school had to ask whether they enabled a respectful conversation and a proportionate response.

The right to challenge and institutional accountability#

Participation includes disagreement. A free, accessible, non-retaliatory route should exist to report an error, seek review of a decision, or request that a feature not be used. The response should state what was checked, what changed, and what further remedy is available. A family should not have to prove with specialist knowledge that a system is wrong. The duty to explain and justify lies with the institution that selected it and the vendor that operates it.

The General Data Protection Regulation provides safeguards in relation to decisions based solely on automated processing. The UK Information Commissioner’s Office explains data-protection rights and accountability duties in accessible terms. In an educational setting, those safeguards mean being able to add context, correct a record, and obtain timely human intervention. They cannot be buried in a long privacy notice or a screen seen only during registration.

Governance that leaves an accountable record#

Every material decision should leave a simple record: the need identified, groups invited, contributions received, change made, remaining risks, and review date. A public summary, anonymised where necessary, lets the community see whether participation had an effect and prevents past disagreement from vanishing when staff or vendors change. Contracts should permit audit, access to documentation, and changes requested by the school. Without those terms, the community can comment only on the visible surface of a system.

UNESCO’s Recommendation on the Ethics of Artificial Intelligence connects governance to human rights, inclusion, and accountability. NIST’s AI Risk Management Framework treats risk management as continuous rather than a one-time approval. In a school, applying those principles requires time, budget, independent support, and a high-quality non-digital alternative. Participation does not legitimise harmful technology; it may properly lead to a decision not to deploy it. Its most valuable outcome is an institution able to listen, explain, and correct before completing a process is mistaken for education.

Institutional and primary sources: United Nations, Convention on the Rights of the Child; European Union, Regulation (EU) 2024/1689, Artificial Intelligence Act; UNESCO, Recommendation on the Ethics of Artificial Intelligence; Information Commissioner’s Office, Information rights for the public; NIST, AI Risk Management Framework.