Shankar Vedantam, known for hosting the popular podcast Hidden Brain, discussed the implications of Big Data at Georgetown University on January 29. (Wikipedia) — *Shankar Vedantam, known for hosting the popular podcast* Hidden Brain, discussed the implications of Big Data at Georgetown University on January 29. (Wikipedia)

Shankar Vedantam, NPR journalist and host of the podcast Hidden Brain, weighed in on the buzz surrounding big data at a Georgetown University Lecture Fund event on January 29. The National Institute of Standards and Technology defines big data as “extensive datasets... that require a scalable architecture for efficient storage, manipulation, and analysis.” As increasingly large amounts of data become integral to countless fields and industries—including healthcare, criminal justice, and all types of business—the demand for a system at the forefront of technology that can sort through astronomical amounts of data to generate insights and conclusions is rapidly expanding.

Interest in “big data” over time (data source: Google Trends, https://www.google.com/trends)

Big Data & Plato’s Cave

Instead of delving straight into the intricate facets of big data algorithms and the controversial implications of technological progress, Vedantam opened his talk with words that seemed to come from an ascetic philosopher rather than an avid student of science. After telling audience members to put away their phones, Vedantam launched into a discussion about Plato’s allegory of the cave, which claims that people who manage to “escape the cave” and see the truth of reality will have a hard time convincing their companions in the cave that the world is more than just shadows on a wall. According to Vedantam, this famous philosophical story has many parallels with the implications of big data. “The people who have access and knowledge to how big data works… they tell us: look, we have managed to escape the shackles of confinement, and we have new ways of seeing reality,” said Vedantam. “Some of us are persuaded, and others of us will say, well, these guys sound deranged.”

Big Data & the Human Brain

To better approach the topic of big data, Vedantam turned to the human brain, which he dubbed, “the most interesting big data machine that sits between [our] ears.” In his view, this burgeoning scientific field reflects the mind’s greatest strengths and weaknesses. It is capable of both “remarkable acts of creativity and genius and insight,” while also being “vulnerable to all kinds of biases and fallibilities and shortcomings.”

Vedantam brought up big data’s use in the criminal justice system to illustrate his point. Specifically, the technology has enabled powerful computer algorithms to outperform judges in deciding whether to put people who are arrested in prison or allow them to go home. Such algorithms can identify which individuals are threatening enough to be held without bail. However, this big data-powered feat comes with a caveat. Citing the computer science catchphrase “Garbage In, Garbage Out,” Vedantam claims that biased data fed to a computer can lead to flawed outputs. Indeed, a machine working with data on people’s criminal records could easily be tainted by the racial prejudice that already plagues our criminal justice system.

Limits of Model-Building

Among Vedantam’s other compelling examples that drew parallels between the brain and big data was a simple question for the audience. The question is as follows:

Can you spot the

the mistake?

1 2 3 4 5 6 7 8 9

As the audience strained to find some inconsistency in the simple arithmetic sequence, Vedantam’s slideshow revealed the repetition of “the,” eliciting several surprised exclamations. “The reason you have not picked up the extra word is not because you’re dumb, but because you’re very, very smart,” the journalist explained. Just like how the brain automatically filters out small details to speed up reading, big data intentionally overlooks small mistakes to boost efficiency, which means that sometimes, it “fails to see what is right in front of it.”

“[Big data] is actually in the business of generating models.”

Vedantam further explained that big data, just like the mind, is “actually in the business of generating models.” Therefore, as we observe reality, “we are actually seeing our internal models of the world rather than seeing the world itself.” Similarly, to process large amounts of information, big data also does not “take in information in a neutral and dispassionate way” and instead opts for the faster track of model-building. The downside, of course, is that the powerful machines inside our heads and inside computers are often susceptible to confirmation bias.

Ultimately, even as big data breaks down barriers and shows us a brand new reality, it may still be vulnerable to humanity’s commonplace, but fatal, biases. In Vedantam’s own words, which put a creative twist on Plato’s ancient allegory, “it’s easy to take the prisoner out of the prison; it’s much harder to take the prison out of the prisoner.”