Moral Implications of Analytics and Cognitive Computing
As we look forward to the Stewards of Change symposium in a few weeks, I have been thinking about the challenges to using advanced analytics and cognitive computing to strengthen human services programs. Of course, I am not alone. Recently, academic and policy journals have published articles wrestling with the potential moral impact of predictive analytics—and by extension, cognitive computing—in the world of human services e.g. Managing the flow: Predictive Analytics in Child Welfare.
These articles raise important concerns about our collective ethical obligation in using advanced analytics to make policy and program decisions in child welfare and other branches of human services that help children and families without impinging on civil rights, and without undermining the professional judgement of human services professionals. As the world of human services becomes both more interoperable and more data-driven, these questions will pose immediate questions to human services agencies. For example:
- Can the use of predictive analytics stigmatize children and families, perhaps exacerbating biases too often seen in HHS?
- Can an over-reliance on analytics replace the professional judgement of trained social workers, or provide an “easy out” for human services professionals to avoid appropriate accountability?
These questions underscore an important fact: as sophisticated and as powerful as advanced analytics and cognitive computing* are, in the end they are simply tools, without an inherent ethical dimension. Like any tool, their ethical “goodness” lies in how they are used. Poorly designed, poorly implemented and poorly managed tools can do real damage, but a well-designed, implemented and managed analytics and cognitive computing program can provide meaningful insight not otherwise available to human services professional. In a well-constructed program, these tools will not replace or second-guess the judgement of human services professionals, but rather augment their judgement, providing insight and advice that can help them make better, more informed decisions, which in turn will lead to better outcomes for clients. In the end, what matters isn’t the tool, but how it is used. In other words, it’s less about the hammer, and more about the carpenter who wields it.
For more specific information about the potential application of cognitive computing and child welfare, please see our blog Doing Good in the Cognitive Era.
*Cognitive systems can learn through experiences, find correlations, create hypotheses, and remember, and learn from, the outcomes.