Hero Image Photo credit: Schloss Dagstuhl – LZI Gmbh
On January 5th, we were invited to take part in a five-day seminar on AI for Social Good at Schloss Dagstuhl – Leibniz Center for Informatics. The seminar brought together AI and ML researchers, and representatives from NGOs working across legal aid, humanitarian response, gender justice, poverty reduction, and child protection.
The week was marked by a contagious collaborative spirit. Rather than a “tech-push” exercise, discussions were grounded in real-world challenges facing NGOs and in a shared motivation to leverage AI tools to meaningfully serve people.
Open Text Analysis
For CHI, the week was an opportunity to explore how AI could support one of our core challenges: analyzing large volumes of qualitative data without losing nuance, context, or ethical integrity.
We teamed up with representatives of 510 Global and Amref Health Africa, and scientists from CIRAD and Media Cloud to use AI to analyze open-text case summaries from child helplines. These summaries describe, in children’s own words, the reasons they reach out for help, often involving overlapping experiences of violence, distress, and mental health concerns.
Questions around AI and Child Protection
Some of the most valuable conversations during the seminar went beyond technical performance. Participants engaged deeply with questions that are especially critical in the child protection sector:
- How do we protect the privacy of children and young people sharing highly sensitive and vulnerable information?
- When is it appropriate to use AI, and when should human judgement take precedence?
- How can humans be meaningfully integrated into AI-supported pipelines, rather than sidelined?
These discussions affirmed that ethical AI is not only about safeguards and compliance, but about intentional design choices: who the system serves, whose voices are amplified, and how harm is avoided.
Looking Ahead
As AI becomes increasingly present in the lives of children and in the systems designed to protect them, being part of these spaces and conversations is essential for CHI. The AI for Social Good seminar offered not only technical insights, but also new partnerships, shared understanding, and a reminder that responsible innovation depends on dialogue between those who build technology and those who work closest to lived realities.
We look forward to continuing the collaborations sparked during the seminar, and to ensuring that as AI evolves, it does so in ways that help us better listen to, protect, and amplify the voices of children.
Tania Cruz Cordero PhD
Data Analyst
