In one of our latest explorations of AI, we harnessed the capabilities of an AI chatbot to craft a narrative from two distinct AFCARS (Adoption and Foster Care Analysis and Reporting System) reports. This exercise demonstrated a powerful use of AI, which is especially helpful for those of us who need to be able to identify trends and interpret data in order to make better decisions.
The Experiment
We prompted Microsoft’s Copilot to generate a narrative from two chosen reports. Here is our prompt:
“Provide a detailed analysis that not only quantifies the differences between these two reports but also hypothesizes potential reasons for the shifts, considering policy changes, economic factors, or other systemic influences.”
Our prompt was detailed, clear and open enough to let the AI make certain assumptions about what would be interesting to child welfare professionals. You might also consider describing an audience such as a child welfare administrator, a policymaker or a CEO (depending on what types of insights you want to see).
The Magic of AI Interpretation
What makes this example noteworthy is the absence of any narrative text within the reports themselves. The chatbot performed all the analysis and inference independently, relying solely on the statistics provided. The AI demonstrated an ability to analyze and infer from raw data interesting conjectures and trends. One notable insight was the impact of the COVID-19 pandemic on foster care numbers—a conclusion drawn without any explicit hints in the prompt.
The Role of Human Expertise
Was all of this analysis completely accurate and coherent? No. But this exercise highlights a crucial point: the AI can offer ideas that you might not have considered in that moment and provides a valuable starting point for exploring ideas about the story that they numbers might be telling. And this is where your expertise comes into play. You can leverage your knowledge and experience to evaluate the AI’s suggestions, determining their validity and relevance. It’s often easier, after all, to refine an existing interpretation than to create one entirely from your own mind.
Conclusion
This experiment underscores the potential of AI in data interpretation and shows us how we can extend our own human expertise rather than providing a substitute for it. The extension of human expertise using AI tools is going to be more and more prevalent in the future. AI chatbots are “idea tools” that can help us explore different approaches to care that we may not have considered before. So why not start now?
Blog post by Paul Epp