The unintended consequences of data analytics

Vast amounts of data and machine learning allow people to see what was once invisible to the human brain, says Eric Berlow, TED senior fellow and complexity scientist.

He likens the deluge of data to food that helps machines get smarter, he tells AXFI Conference attendees in Minneapolis.

But humans have to be smart about how they use data, machines, and artificial intelligence.

“Trusting it as a value-free black box is dangerous,” Berlow says.

He cites two examples of potential unintended consequences:

1. Amplifying crowd behavior. Overuse of the most popular articles on a news website leads to the most important stories not getting enough attention.

2. Increasing inequality. Using certain data sets could end up being discriminatory to certain populations.

So how can credit unions combat these and other unintended consequences?

Look at the data, Berlow suggests. Humans are good a differentiating and can spot errors or bad connections.

“Just looking at the data is really important,” he says.