Maligned Pentagon AI program ‘GAMECHANGER’ might inspire Congress to change its own game

A recent article in the Intercept portrayed the Pentagon’s GAMECHANGER program as an embarrassing admission of bureaucratic ineptitude. That kind of clickbait headline exacerbates the very problem it criticizes. There’s little need for debate about the crushing burden of policy cruft the Department of Defense (DoD) operates under — even the DoD itself has long since ceased being defensive about it — but mocking the department’s efforts to increasingly automate the application of its policies and budget only serves to demoralize public servants fighting for change. 

GAMECHANGER uses AI derived from free, open-source natural language processing libraries to compile over 15,000 policy and budget documents governing how the Pentagon and the services operate. As the article notes “the annual Department of Defense budget legislation includes hundreds of revisions and limitations telling the Pentagon what it can and cannot do.” GAMECHANGER helps make sense of all those provisions.

Navigating the complexity and sheer volume of Pentagon policy — equivalent to 100 copies of “War and Peace” — slows everything from acquisitions to hiring to logistics to combat operations. According to Stuart Wagner, one of the co-founders of the project, “Warfighters each day are required to assess laws of war against planned operations. If the time to understand the policies exceeds the time window, the operation gets canceled.” There are plenty of reasons an operation should be canceled, but failure to understand our own rules should not be one of them. GAMECHANGER ingestspolicy as data, then accelerates policy analysis, contradiction discovery, and authoritative guidance to expedite defense decision making. 

In a LinkedIn post, Wagner documented the justifications to develop GAMECHANGER. First, policy discovery and analysis take extensive resources. “A new DoD agency leader asked what his team’s authorities and responsibilities were. His team spent nearly a year finding and reading federal documents to answer,” illustrating the very problem GAMECHANGER ultimately solves. Second, policy drives security classification determinations, and the inability to rapidly determine the classification often results in a default overclassification. GAMECHANGER helps with this as well, but negative press coverage like that from the Intercept will only serve to strengthen the internal opponents of efforts like these. By dismissing the team’s ingenuity and dedication, the Intercept shapes the DoD to be more like the institution it derides. 

There is no doubt that a far better solution to the DoD’s troubles would be a bold, sweeping rationalization and simplification to increasingly automate the application of its policies. But the Pentagon is not unique in this regard. In my recent book, “Recoding America,” I tell of a claims processor handling unemployment insurance claims for the state of California who is still considered a beginner after 17 years in his department, so complex and arcane are the policies that govern the unemployment insurance program. And whether its state labor agencies or the Pentagon, this complexity is not entirely self-inflicted. Policy descends on these institutions from Congress and state legislatures, the White House and governor’s offices, and judicial decisions. Rules are always added, almost never removed, accumulating over the years like so much clutter. Clearing out that clutter will require much more than artificial intelligence; it will require the sustained attention of Congress. 

Perhaps Congress can be inspired by GAMECHANGER to change its own game. Instead of AI to simply interpret millions of pages of documents, our lawmakers could use AI to boil those millions of pages down, suggest dramatically streamlined versions, and repeal the clutter en masse. The authors of GAMECHANGER may have created more than an innovative way to manage a dangerous mess. They could inspire our leaders to clean it up. 

Jennifer Pahlka is the founder of Code for America and former US Deputy Chief Technology Officer for the Obama administration, where she helped found the United States Digital Service. She served on the Defense Innovation Board under Presidents Obama and Trump and is the author of “Recoding America: Why Government Is Failing in the Digital Age and How We Can Do Better.”