big_data_003
TechTank

Five Ways the Government Wants to Use Big Data (Beyond the NSA)

Joshua Bleiberg

Much of the conversation around government uses of big data has centered on the controversial NSA programs that crawled social media and telephony meta-data.  Many expected the recent report from the Obama administration to address concerns about the use of big data analytics by intelligence agencies.  The report punts on that hot topic, which disappointed many observers hoping the administration would roll out a government wide privacy policy for big data.  The authors write, “This report largely leaves issues raised by the use of big data in signals intelligence to be addressed through the policy guidance that the President announced in January.” Big Data: Seizing Opportunities, Preserving Values does offer broad coverage of the different ways the federal government would like to use big data.  Here are five of the most interesting proposals for the use of big data by government.

Do Not Pay Portal

The Department of the Treasury has created a “Do No Pay Portal”.  The intention is to create a single website where government workers from every agency can access information to prevent fraud, errors, or identify ineligible participants.  According to an inspector general’s report the IRS handed out nearly $4 billion in fraudulent tax refunds over the past year.  Currently the portal only includes government data sources but Treasury would like to supplement their efforts with commercially available records.  Comparing records with big data analytics could make it easier to crack down on tax cheats. 

Continuous Evaluation of Insider Threats

A number of troubling events including the Washington Navy Yard shooting, Fort Hood shooting by Major Nidal Hissan, and the WikiLeaks disclosures have called into question the procedures for providing access to facilities and documents to government employees.  The Army is testing a program called Automated Continuous Evaluation System.  The system analyzes government, commercial, and social media data.  The program revealed 21.7 percent of those in the pilot had not disclosed important information like serious financial problems, domestic abuse, drug abuse, or allegations of prostitution.  In about 3 percent of cases the charges were serious enough to result in the suspension or revocation of their clearances.

Helping Students Learn

The Department of Education has incorporated big data into the National Education Technology Plan.  In the education sector, developers use data mining to apply lessons for learning specific competencies.  The Ed Department hopes to one day use similar big data tools to analyze courses, schools, and even entire districts.  These plans are in the early stages but represent an exciting new field within education.

Doing Away with Fee-For-Service in Healthcare

The U.S. healthcare system relies on a “fee-for-service” model where medical professionals are paid regardless of whether the treatment was effective.  This creates an incentive for care providers to submit their patients to unnecessary treatments that will have little or no health benefits.  The Affordable Care Act helped to ease the transition to electronic record keeping that would allow big data analytics to flourish.  Long-term the government could use big data analytics to inform how it reimburses doctors through Medicare and Medicaid.

Tracking Illegal Activity on the Deep Web

The Memex program developed by DARPA enabled federal law enforcement officials to further their investigation of human trafficking in the United States.  Memex gathers information on the “deep web” or pages that are not indexed by popular search engines.  Analytics combine together information from the surface web, deep web, and law enforcement databases to detect criminal activity.  Although this is possible without big data tools Memex saves officers time and resources.  The government would like to use pattern analysis to identify all criminals that rely on the deep web to communicate.

Though none of these policies are as controversial as PRISM they each offer their own benefits and tradeoffs.  Big data presents unique problems because it is anathema to privacy.  To make big data useful requires the sharing of personal information.  Addressing this core problem will require balancing the competing values of making government more efficient and respecting the privacy of U.S. citizens.

Author

J

Joshua Bleiberg

PhD student, Vanderbilt University. Former Research Analyst, The Brookings Institution.