Do you ever feel as if the internet is following you?
It is — at least some segments of it.
Ads on Facebook and other websites will appear freakishly fast after a related Google search and stay in your line of vision until you start to search for something else. Some people call them “stalker ads,” said Paul Cimino, a tech entrepreneur and data strategist, who led a discussion on digital communications at Fordham University’s monthly CIO Roundtable.
The ads, Cimino said, are driven by machine learning. Large and tremendously complex sets of data are fed into computers, which react quickly to funnel ads to you on other sites.
The process is called retargeting or remarketing, and it “is driving more than half of all display ads right now,” Cimino said.
Cimino’s remarks, delivered to the group of technology officers and professionals, came during a conversation that ranged from the future of online communication to the dilemma of ad-blocking software.
The growth of data and “bot” technology has led to increasingly sophisticated uses, of which retargeting is one, he said. Data are gathered through people’s internet habits, such as visiting three pages on a given website or filling out a portion of a registration form.
Cimino described his own internet practices as an example.
Machine learning is taking humans out of the equation in this segment of internet advertising, he said. “It’s basically going to be robot programs battling each other without human intervention” — though humans are still needed to set the parameters and desired outcomes.
That human involvement has the potential to skew the results of the data, said Timothy Malefyt, who did not attend the roundtable but who is among the marketing faculty experts at the Gabelli School. Algorithms are created by programmers, he said, and their output may affect programmers’ biases.
Big data “provides an illusion of science as rational and intellectual,” Malefic said, but “it is ultimately highly social.”
Looking at technology more broadly, Cimino sees tremendous potential for innovation in the next 20 years as wearable computing, infrastructure advances and human imagination converge.
“We essentially have a mainframe on our hip, with the cloud, but it doesn’t really do a lot for us,” Cimino said. “But I think in the future it’s going to be profound.”