A chatbot informed by children’s experiences
The development of Advisy was informed by a global consultation with child human rights defenders conducted by Child Rights Connect in early 2025. The survey gathered the views and experiences of 139 children from 20 countries across 6 regions of the world, available in six languages.
The findings painted a clear picture of the challenges CHRDs face online: while the vast majority are digitally active — using social media, online platforms, and digital tools as part of their activism — 55% reported having experienced at least one digital security incident, and 42% had been victims of a serious security breach.
The survey also found that while children take steps to protect themselves, they need better information, stronger education, and more accessible tools. Advisy was developed as a direct response to this need. The full findings of the survey are available here.
Developed with children, for children
Child participation was at the heart of Advisy’s development from the very beginning.
Since 2024, two successive cohorts of Child Rights Connect’s Children’s Advisory Team (CAT) have been involved at every stage of the process. The child advisors first received capacity building on digital protection for child human rights defenders, which provided the foundation for their meaningful engagement with the chatbot’s development. From there, they participated in a series of dedicated meetings alongside the technical consultant responsible for building Advisy, who explained the technical background in accessible terms and adjusted the chatbot live in response to the children’s feedback.
The child advisors shaped Advisy’s identity in fundamental ways — it was they who decided on the persona, the seagull character, the name, the introductory message, and the visual design including colours. They also tested the chatbot extensively and rigorously, checking the boundaries of its scope by asking it questions outside its purpose — such as how to bake a cake or how to complete their homework — and assessing how it handled these situations. This testing was conducted in multiple languages, including the mother tongues of the child advisors, ensuring that Advisy’s safeguards and limitations were verified across different linguistic contexts.
Our commitment to safety
Child Rights Connect takes the safety of child human rights defenders seriously and has approached the development and deployment of Advisy with the same safeguarding standards it applies to all its activities involving children. The entire development process was guided by the UNICEF Guidance on AI and Children 3.0 (December 2025).
A comprehensive risk assessment was conducted in line with the organisation’s safeguarding policy, identifying risks and proposing concrete prevention and mitigation measures, all of which have been put into practice. Internally, a monitoring and review framework has been established to ensure the chatbot continues to operate safely over time, covering three-weekly human monitoring of conversations, six-monthly reviews, annual assessments, and a set of immediate actions that can be triggered at any time should a concern arise.
A child-friendly safety guide — “Before you talk to me“ — was developed to ensure that children can make an informed decision about using the tool and know how to protect themselves before they start.