We’re proud to share the outcomes of an innovative collaboration with the University of Toronto’s Data Sciences Institute (DSI), a partnership that brought together child welfare expertise and cutting-edge data science in service of the children, youth, and families we support every day.
About the Initiative
Through the DSI’s Summer Undergraduate Data Science (SUDS) Scholars initiative, CAST partnered with a team of graduate and undergraduate researchers supervised by Dr. Shion Guha, working in close collaboration with CAST staff. Critically, this work was supported by the Child Welfare Institute and Information Management teams, whose practice expertise and data stewardship were central to the effort. Together, the team explored how responsible, human‑centred data science and AI tools could be applied to real child welfare challenges, including the complex task of identifying patterns within narrative casework documentation.
Applying Technology Responsibly
The research demonstrated the potential of large language models (LLMs) and computational tools to thematically distill narrative case data, supporting caseworkers by surfacing patterns and insights to inform decision‑making without replacing professional judgment.
As one student researcher reflected, “Instead of only being concerned with what my tools do, I was able to learn how to apply my tools thoughtfully and carefully for a cause with real impact on a vulnerable community.”
A True Collaboration
What distinguished this collaboration was its deeply human‑centred and participatory approach. CAST staff partnered closely with the research team throughout the co‑design and analytical phases, ensuring the work remained firmly rooted in the realities of child welfare practice.
One team member noted: “The collaborative dynamic between CAST and our lab was great, a lot of the patterns and insights we found wouldn’t have been possible without the CAST team’s insights and suggestions throughout the process.”
Impact on the Next Generation of Data Scientists
For the students, the experience transformed how they think about data science. “This experience helped me understand the importance of critically evaluating algorithm performance in real-world systems,” shared one researcher, “and made me internalize the potential impact of these models on a person’s life.”
Looking Ahead
For CAST, this partnership reflects our ongoing commitment to evidence‑informed practice, innovation, and the responsible integration of technology—always with people at the centre. This work also reflects the Child Welfare Institute’s continued leadership in advancing the co‑development of practice‑informed tools that meaningfully support child welfare services. As one researcher noted, “I hope through this work, future researchers will be encouraged to develop child welfare tools with a human‑centred focus, prioritizing the downstream user throughout the process.”
We are grateful to the University of Toronto’s DSI, the SUDS Scholars program, and the research team for their dedication, care and curiosity. Collaborations like this one remind us that when we bring the right people together around a shared purpose, the potential to improve outcomes for children and families is limitless.
