Big data can provide a tremendous societal benefit, and there are many examples of analytics being used for social good. Yet, there is immense concern that data can be manipulated and directed against the very people that are already at a disadvantage, such as lower-income families and youth. Civil rights advocates argue that automated decision-making could worsen discrimination without proper oversight.
The Annie E. Casey Foundation is a private philanthropy dedicated to developing solutions that strengthen families by building economic opportunity paths. They recently released a white paper with their suggested best practices for data usage to effectively inform social programs and policy. BCT Partners, a national consulting firm, who has a mission of providing insights about diverse people that leads to equity, contributed observations included in this paper.
The Casey Foundation recommends four guiding principles in using “analytics for good.” They include expanding opportunity, providing transparency and evidence, empowering communities, and promoting equitable outcomes.
Principle 1 - Expanding Opportunity
The first principle is to use data to exponentially better the lives of children and families and not "merely improve the status quo incrementally.” The recommendation is that rather than data usage identifying risk or preventing worse harm to individuals and communities, analytics should identify new opportunities for progress. For example, modeling could help spot talented youth, create career path recommendations, and match them to supportive organizations. In another actual example, Los Angeles used machine learning to access millions of criminal records to find those that had marijuana offenses. The city then expunged those charges from criminal records giving many individuals a proverbial clean slate.
Principle 2 – Providing Transparency and Evidence
Decisions that significantly impact people’s lives must be made with full transparency. Government agencies or businesses that use data for policymaking decisions should provide supporting evidence of how and why they arrived at their conclusions. Local and state task forces offer an excellent resource for understanding how to best engage communities in the use of advanced analytics and assist in the development of policy frameworks that regulate usage. “One well-known effort that engaged the public and provided full transparency in the development process is the Allegheny Family Screening Tool used in Allegheny County, Pennsylvania, to screen calls made to child protective services.” Because the development and evaluation took place in public view, constituents had the opportunity to vigorously debate and provide feedback throughout the process, thereby ensuring the tool provided the desired outcomes.
Principle 3 – Empowering Communities
Communities should have access to analytics to hold institutions and leaders accountable. And that data should be employed to remove discriminatory practices and systematic barriers against equal opportunity. More often than not, information is used against communities of color. For example, many police departments use predictive algorithms to target individuals and neighborhoods for surveillance. Yet, they have resisted using the same predictive technologies to examine how citizen complaints on officer conduct might prevent unnecessary use of force by police. Analytics should be used for the betterment of communities and not to justify more prejudicial practices.
Principle 4 – Promoting Equitable Outcomes
“New tools and models should explicitly promote, and be judged against their ability to achieve, more equitable outcomes for historically disadvantaged groups.” Investment decisions would then be made based on whether analytics projects correct inherent biases used to support discrimination against people of color. And advanced analytical tools should be tested and shared before implementation with the intent of rejecting those that worsen the negative treatment of people of color.”
Conclusion
In summary, there are endless possibilities for ways that data can improve outcomes for disadvantaged people. The social sector is just beginning to tap into that knowledge with the amount of available information continuing to grow. However, all relevant constituencies must work together to ensure that this private data is protected from misuse and the end results are for public good.
Download the full report here
To read more BCT blogs, go here
*Quotes were taken directly from the Casey Foundation’s research paper