Can Politics Be Debugged?

It’s a Silicon Valley approach to some of the most pressing political problems facing the U.S. today.


How do you fix U.S. politics?


It’s a question that seems almost too daunting to even begin to answer. But a team of developers and designers at the organization Debug Politics had an idea: Why not apply Silicon Valley’s favorite way of generating new ideas to America’s broken political system–and combat the conservative policies of President Trump in the process?

That’s right, they had a hackathon. And not just one–the organization has now hosted three Debug Politics hackathons, two in San Francisco in late 2016 and one in January in New York, with another upcoming this spring in Los Angeles.

At the most recent Debug Politics event in New York, the offices of the cult mattress startup Casper buzzed with urgency and thrumming Macbooks. It was just minutes before the final product showcase, where 20 teams, generally composed of four to five members, would demo their ideas before a panel of judges from VC firms, design agencies, media, startups, and the Clinton campaign. Jesse Pickard, one of the event’s organizers and the CEO of the education app Elevate, stood up. He wore an American flag cap, backwards.

“November 8, election night, was actually my birthday,” he said. “It was a very sad, sad birthday. We never got a chance to eat the cake.”

Devastation turned into motivation, Pickard described, and he and a few of his designer and developer friends discussed what to do next–a conversation had by many across the country. “We talked about this clear choice that we had to make. There were two very obvious options. The first was to continue to recede into our echo chambers. Post angry tweets. Post a huge long thing on Facebook. Commiserate. Sulk,” he said to the crowd. “And then there’s option two, which I’m a far bigger fan of. Option two is getting off of our asses and using the skills that we have in this room to actually bring about positive change.”

He’d told the same story two days earlier when the hackathon opened, and his words were a call to code; about 40 teams made it through the entire weekend of work, and though there was only time for 20 to present, Pickard said it was the best turnout they’d had so far, with 350 participants and observers over the course of the weekend.


“We’ve probably dropped the ball”

Civic technology, or technology oriented toward helping local, state, and federal governments, isn’t new. But the election of Donald Trump has spurred interest from techies looking to use their skills, often employed in jobs that provide cushy perks, in a more positive way.

“We get paid a lot, we have these wonderful opportunities at work where we have free food and everything like that, and it’s easy to forget that there’s a broader community out there that is struggling in different ways. For the past couple years we’ve probably dropped the ball a little bit and gotten too comfortable,” Pickard says. “I think we need to dedicate some of our time each week to bettering our community. I think it’s an obligation. It’s not about if we can or should help. We need to.”

Clearly the hundreds of citizens who showed up to the Debug Politics hackathon believed so, too. Over the course of 48 hours, the 40 teams built platforms to combat fake news, encourage participation in local politics, examine bias in reading habits, and decipher and annotate incomprehensible laws. One group even made a tool that allows you to enlist your Twitter account in an alt-left fake news bot army. The winner, Second Opinion, is a Facebook Messenger app that will evaluate a news article’s trustworthiness and find alternative viewpoints on the same topic. The runner-up, Stand Up, is a platform for making coordinated calls and visits to your representatives.

The challenges that face these projects are considerable. On the second day of the hackathon, Matt Stempeck, the director of civic technology at Microsoft who served as the director of digital mobilization for the Hillary Clinton Campaign, had given a short lecture and answered participants’ questions. Stempeck says that one of the biggest issues with civic technology lies in the limited amount of data accessible to developers–something that many of the platforms built at the hackathon would need in order to succeed. While the Obama administration has done much to provide well-formatted government data to the public, there’s still a lot of data that isn’t easily accessible. Not only that, governments are slow-moving and often require individualized solutions, making them difficult clients for startups looking to improve infrastructure and bureaucracy.

One issue with putting a lot of eager coders and designers in a room and asking them to debug politics is that they may lack context about what other people in the civic tech space are already working on. That’s why Debug Politics partnered with the local organizations New York Immigration Coalition and Asian Pacific Public Leaders, which each suggested hackathon projects that would be useful for their work. None of the participants who had indicated they were interested in working on one of these suggested projects had ultimately chosen to build them, which Pickard attributed to the fact that neither of the organizations had representatives there on the first night of the hackathon to give participants guidance. It underlines the inherent tension with hacking politics–a complex network of local and national groups whose problems are varied and often require smaller, organization-specific solutions.

One team did choose to partner with an existing organization–the News Literacy Project–in order to redesign their website. This kind of pro bono work is a great way for designers and developers to begin making a difference in the political sphere without starting something entirely from scratch, especially when they may not have any institutional knowledge about civic technology.


Still, technology is really good at improving systems of information and communication. For young people unsure about how to make a difference in the political system, Stempeck believes tech is a good hook and can help give people disillusioned with the current system hope that the government can serve the people again. “Tech gets you in the door, and then you stay for all the other reasons, like fixing democracy,” he says.

Mapping The Hidden Paths Of Power

One team in particular took the hackathon’s name to heart. Composed of engineers Peter Dixon-Moses, Riley Martinez-Lynch, Steve McCarthy, Arnaud Sahuguet, and Margo Smith, the team started a project called Influenza, a nonpartisan tool to help identify problematic relationships in the U.S. political system. “There is this maze of influence, and we’re trying to make sense of it,” says Arnaud Sahuguet, director of the Foundry at Cornell Tech and former CTO of NYU’s Governance Lab, who presented the project at Debug Politics.

Here’s how it works. You start by identifying what the team calls a “bug”–that could be nepotism, a conflict of interest, pay-to-play–basically, a questionable series of relationships that could impair the judgment of an elected or appointed representative of the United States.

Take for example this rather complicated scenario: If an elected official were to appoint an agency head who runs the Department of Justice, which is investigating a company to whom a company owned by the elected official owes money, that would be considered a conflict of interest. It’s a convoluted relationship to wrap your mind around–so why not let a computer parse it for you?

This series of relationships between elected official, agency head, agency, company under investigation, and company owned by the elected official can then be translated into computer code. (As of right now, anyone using the tool has to identify a “bug” themselves; the tool currently lacks any accessible user interface that non-coders could use.)

Using this code, you can then search the Influenza database, which contains encoded information about elected representatives and their relationships to each other and to corporate entities, for that exact relationship. The tool shows you every instance of the “bug” you’re searching for throughout the entire U.S. government. Eventually, users might be able to simply choose a type of bug and then see all the instances of it, but there’s a lot more engineering to be done before that point. At its current stage, Influenza is more of a framework, with broad potential applications in the future.


In this case, the database brings up Donald J. Trump’s relationship to proposed Attorney General Jeff Sessions, Deutsche Bank (which is under investigation by the DOJ), and Trump Holdings (which owes Deutsche Bank money), all without searching for any of those names specifically–making it an objective, nonpartisan way to search for problematic relationships within the government.

“I’m not going to tell you this congressman is doing something bad,” explains Sahuguet. “But we can agree on what kind of behavior should not be present in U.S. politics. It’s not just my opinion and my bias and your opinion and your bias. We have a common understanding and agreement about what is a feature and what is a bug. And then we can decide what we want to do about it.” It’s a scientific approach to political problems.

While Influenza’s database is composed of data sets from, OpenSecrets, ProPublica, and LinkedData, the team had to encode some of the corporate data manually because open, public data sets don’t exist yet for that kind of information. But Sahuguet believes that this kind of data exists at the DOJ or at a newswire like Bloomberg–it’s just a matter of tapping into it to make the tool more robust.

Sahuguet says that Influenza, which is currently available on Github, is the kind of tool that could be a dream for data journalists, though it’s far from being accessible on a website (though BuzzFeed’s investigative team is working on something similar, using crowdsourced data). He’s already taken Influenza to his colleagues at Cornell Tech and hopes that students will be interested in working on it, and that one day a tool like this will be commonly used to assess if political appointees are qualified for a post. He even envisions Influenza constantly running queries on the blockchain where censors can’t reach it, and sending notifications when it finds a problematic relationship.

It’s a lofty idea, grounded in data—and one of the best to come out of the Debug Politics hackathon.

In the end, the event exposed both the challenges and opportunities that lie in applying the tech world’s idealism to the government. There are serious practical and cultural roadblocks to applying a “move fast and break things” attitude to the political system. For all of the talk that it’s broken, breaking it further isn’t actually going to help. When lives and livelihoods are on the line, there’s no room to fail.


Still, Pickard says this drive to fix bugs is what has inspired so many designers and developers to hunt for solutions to political quandaries. “At the end of the day, I think tech people like to solve problems—more specifically they like to fix broken things,” he says. “I think after the last year with the crazy election and then the result, tech people are starting to think of politics as one of the most broken things to fix out there.”

About the author

Katharine Schwab is the deputy editor of Fast Company's technology section. Email her at and follow her on Twitter @kschwabable