#AskTheExpert invites the Eyebeam community to ask an expert questions pertaining to their field. This week we opened up questions to Surya Mattu, an artist, engineer, journalist, and Eyebeam alum. His practice combines art, investigative journalism, engineering, and creative technology. Currently he’s an investigative data journalist at The Markup, but previously was a contributing researcher at ProPublica, where he worked on “Machine Bias,” a series that highlighted how algorithmic systems can be biased and discriminatory. He’s also Knight First Amendment Institute’s first Resident Technologist.

 

@munigram: How much will algorithms predetermine a child’s educational path (gifted programs, college acceptance, STEAM skills) in the future?

Surya: The short answer is a lot, though probably not in obvious ways. The biggest challenge in answering that question accurately is that most algorithmic systems are black boxes that we can’t interrogate because they are protected as a company’s trade secret. While the algorithms occupy the spotlight in public discourse, the reality often is that they are just a convenient scapegoat for a system that is cripplingly under-resourced. I highly recommend Virginia Eubank’s Automating Inequality to see how this has played out in other parts of society.

Additionally, these things don’t always play out in intuitive ways. For example, a recent Washington Post article exposed how college admissions officials rank prospective students before they apply. It highlights a growing trend, not only do algorithmic systems perpetuate systemic biases that are already prevalent in society, they also obfuscate more predatory practices.

@huuugeposer: I’m always curious about what algorithms are set in place and how I can work with them. It ends up being a battle of trying to outsmart the algorithms to use social media in the way I want to be seen. What ways can we find out confirmed algorithms and use it to our advantage?

Surya: Sadly, I don’t think there are any meaningful ways to outsmart algorithms because as soon as a technique becomes public the platform usually updates their software to block it. 

Having said that, the ‘Developer Tools’ on browsers is a good way to poke around and see what information the platforms are sending and receiving from their servers. Developer Tools are tools that are built-in to all the popular browser that allow a user to dig a little deeper into the website they are visiting. It lets you see the raw HTML and Javascript that is running on a page, as well as all the network requests which can be incredibly useful.  This course provides a simple introduction to those tools. I use the developer tools all the time, usually when starting a new investigation.