Data sharing initiatives should be built on accountability and security, expert says

Breaking down silos and enabling data sharing across agencies has long been a priority for state and local governments eager to optimize actionable insights and service delivery. But new guidance suggests that data consolidation should be done through the lens of privacy and security. 

When adopting a centralized data infrastructure, government and agency leaders should consider that “a number of technical considerations during implementation will impact the capabilities, and thus the consequences, of a combined data system,” says a technical guide recently published by the Washington, D.C.-based nonprofit Center for Democracy and Technology. 

“As more agencies contribute to consolidated datasets, the potential for errors expands …. [which] has the potential to cause enormous harm to people caught up in those government actions, which could include inappropriate benefits stoppages, false arrests and the overall degradation of government services,” according to the guide. 

Such risks “may not comport with the expectations and consent of those who provided the data, violating their privacy and degrading trust in government agencies. The potential for these wide-ranging and damaging impacts requires careful attention to and governance of government data sharing programs,” it says. 

For state and local governments, their management of shared data moving forward is particularly pertinent amid mounting demands from the federal government to obtain data that it has previously not had access to, Hannah Quay-de la Vallee, senior technologist at CDT and author of the technical guide, told Route Fifty.

President Donald Trump, for instance, signed an executive order last year directing agency leaders “to ensure the federal government has unfettered access to comprehensive data from all state programs that receive federal funding, including, as appropriate, data generated by those programs but maintained in third-party databases.” 

The order and continued attempts from the Trump administration to access states’ driver’s license and voter roll data have stoked concerns about data privacy when it comes to government data consolidation, Quay-de la Vallee said. Indeed, recent survey findings found that 74% of Americans are worried about governments’ ownership and storage of their data. 

The CDT guide suggests that agencies consider not just the method through which they break down data silos, but how they will prepare and protect data before centralizing it. 

The guide suggests that agencies can take a warehouse or federation approach to data consolidation. Warehousing is the process by which “all shared data is stored in a centralized location that every party has access to,” and federation refers to a “sharing agency providing access to their own version of the data to the receiving agency,” according to the guide.

Indeed, “it doesn't really matter how you structure [data],” Quay-de la Vallee said. “What matters is the movement of data between agencies and between entities,” she added. 

“The thing that matters about either [warehouse or federation] technical approach that you take is going to require certain kinds of technical interventions,” Quay-de la Vallee said, adding that “regardless of how you have done [data consolidation], how are you making sure that it works as intended?” 

For instance, the guide recommends that agency leaders consider who ultimately controls the data to “ensure effective responsibilities and roles for building the sharing infrastructure.” Under the warehouse approach, one agency assumes the primary responsibility for maintaining and coordinating data sharing, while under the federation model, each sharing agency is accountable for their own data and how receiving agencies access it.

Formal procedures, such as data sharing agreements, are one way for agencies to identify and establish expectations regarding how data will be securely shared and protected by receiving agencies, according to the guide. 

Another technical consideration for government agencies to consider is how they standardize data quality, definitions and matches, Quay-de la Vallee said. 

Such details can be as small as reconciling how separate agencies format a person’s date of birth in their datasets, but “trying to use data with these disparate schemas without clarifying can introduce numerous errors” in the data sources and the outputs they produce, the guide says. 

To reduce the potential for data entry or formatting errors when consolidating information, data standardization and matching efforts are “where documentation does a ton,” Quay-de la Vallee said. 

Agencies can create data matching frameworks where all involved parties can reference a shared document that details how data elements will be, for example, defined or denoted within a central database, she said. Beyond matching frameworks, agencies can draft memorandums of understanding or contracts, she added. 

Such procedures “must include a way to receive information about errors as well as steps to correct the error in a given case and, if applicable, avoid the error in the future,” the guide reads. 

“The further away [erroneous] data moves from that sourcing agency, the harder that problem gets to fix,” Quay-de la Vallee said. Additionally, she said, one department could be “causing problems for other agencies down the line” if staff have to track down an error within the sourcing agency and if that error appeared in receiving agency’s data as well. 

Preparing agency-level data for consolidation is an especially critical consideration for governments as artificial intelligence tools become more ingrained within their processes and operations, Quay-de la Vallee said. 

The tech can, for instance, assist with data analysis across government programs to identify fraudulent benefits claims to boost overall program integrity, Quay-de la Vallee said. But its use cases underscore why data cleansing, framework and documentation efforts are necessary before implementing AI tools that could be used to influence major decisions impacting residents’ lives, she said.