Meaningful Child Participation
Using AI to accelerate global research without removing human judgement. Exploring whether an AI-enabled research assistant could reduce manual research time for Meaningful Child Participation.
Context
Save the Children wanted to strengthen Meaningful Child Participation: involving children as genuine partners and contributors rather than simply beneficiaries. The global evidence base was large and fragmented across charities, brands, platforms and movements. Teams needed to understand strong participation and apply it to their own work, but specialist staff time was being spent on manual desk research.
My Role
As Innovation Lead, I scoped the opportunity, found technical capability and led delivery and internal adoption. I first explored what we could do internally, then concluded specialist technical and research expertise was needed. I recruited and managed an external developer and researcher team, bridged them with internal experts, coordinated feedback and validation, defined trust and usefulness requirements, and supported training, adoption and handover. I was the project lead and client-side owner, not the technical developer.
Problem
How could Save the Children make global best-practice research on Meaningful Child Participation faster and more useful, while ensuring children are treated as genuine contributors and evidence is interpreted responsibly?
Approach
We used a hybrid human + AI research model. An external specialist team used AI to accelerate evidence gathering, categorisation and pattern recognition across publicly available global research. Researchers and Save the Children experts checked quality, context and usefulness. Before wider use, we considered accuracy, safeguarding, AI governance, diversity and inclusion, expert validation, usefulness and human oversight. The work reviewed 130+ organisations and 350 participation activities across charities, consumer brands, digital platforms and collaborative movements. It produced the R.A.D.I.O. framework: Rights (ethical, respectful and safe); Active Role (children influence real decisions); Depth (meaningful, not tokenistic); Inclusion (diverse backgrounds represented); Outcome (involvement leads to tangible change). A practical playbook covered workshops, focus groups, child-review panels, co-creation and youth advisory structures.
Experiments / Evidence
The research identified patterns in how organisations share power with children and where participation becomes tokenistic. It surfaced an organisation-level guidance gap: existing models often help teams run individual activities but offer less support for assessing participation quality across an organisation. The work developed five quality dimensions and a practical playbook, with expert review and human judgement retained throughout.
Outcome
The research assistant was estimated to release approximately 10–20 hours of desk-research capacity per week. I supported training, guardrails and adoption so the internal team could use the capability independently. The tool and research subsequently supported work within the UK Impact team, including policy development around Meaningful Child Participation.
What I Learned
I recognised that we did not have the specialist capability to build this properly ourselves. Innovation leadership can mean defining the problem, bringing in expertise, helping it work with the organisation and making sure the capability has a home. Automate the research workload, not the judgement: AI helped search, structure and identify patterns; internal experts decided what those patterns meant and whether recommendations were appropriate.
