A GOP donor who once had ties to IBM is the man behind the millions of Twitter bots President Trump counts on as followers who could be employed to target voters with misleading or fake news on social media.
In order for it to work, the scheme relies on the quiet guidance of Robert Mercer, a reclusive Republican mega-donor and staunch Trump supporter, sources told the Daily News.
Mercer, who during the campaign with help from his daughter Rebekah, played a below-the-radar — but integral — role in raising money for Trump and connected him with critical figures like Stephen Bannon.
He’s a former computer scientist who, decades ago, helped build the technology at IBM that the computer giant would later use to create its Watson super-computer. He also helped other companies like Apple form the basis for its Siri platform.
Trump, meanwhile, has accrued among his 30 million Twitter followers — 15 million of which are actually bots that experts have told the Daily News could be “weaponized” to spread fake and misleading news stories that favor the White House or distract from the scandals it now faces.
But with the subtle introduction of advanced technology, by individuals familiar with artificial intelligence, their effectiveness as it pertains to targeting users and interacting with them could escalate rapidly.
“Fooling humans into doing things in the electronic realm turned out to be really easy,” said Simon Crosby, the chief technology officer at a firm called Bromium, pointing to well-documented campaigns that are thought to have influenced the 2016 presidential race and the Brexit referendum in the United Kingdom.
“But with a few extra steps that seem available in the right circles, it could get even easier,” he added.
Some of the technology that’s part of Watson, explained Crosby and two other computer scientists who preferred to keep their identifies private, can “quickly build, test and deploy bots or virtual agents across mobile devices or messaging platforms to create natural conversations between apps and users.”
That technology is even built to understand a user’s personality, tone and emotion and interact naturally with people.
“You have arbitrary and ridiculous information spread very quickly, and now to targeted user more susceptible to believing it and spreading it, and we now know that it influences people. We saw it in the election, with Brexit,” he said.
“Ultimately, the problem is that anybody can talk to anybody, especially if ‘anybody’ is a bot on Twitter or Facebook who knows a lot about you.”
All of the scientists were keen to point that such efforts would not be employed by Watson, an entity owned and patented by IBM.
Rather, they indicated, it would appear to make sense for a man familiar with its inner workings to draw on his own expertise to help hone a bot-operation for a political candidate.
While pretty much any one computer scientist with the right motives could help someone in the Trump camp do such a thing — most AI technology at this moment is open source and freely available online if one knows where to look — the proximity of Mercer to the Trump camp, in particular, raises questions.
Long before making his fortune at Renaissance Technologies, his current hedge fund, Mercer, 70, worked at IBM, where his name appeared on several patents and on in many publications and studies about “Brown clustering,” which the computer giant used to create its Watson Artificial Intelligence systems.
Mercer did not respond to questions about his involvement with the Trump campaign.
Mercer, who helped get Breitbart News off the ground years ago with millions of dollars in donations, also invested at least $5 million in a firm called Cambridge Analytica, a small tech operation “that mines online data to reach and influence potential voters” and “uses secret psychological methods to pinpoint which messages are the most persuasive to individual online viewers,” The New Yorker magazine reported earlier this year.
That kind of technology has already been employed in campaigns candidates whom Mercer supported, including Trump’s, and could easily be used in future efforts to target voters on social media, Crosby said.
“We’re in deep trouble in the sense that it’s extraordinarily easy to automate a generation of arbitrarily absurd and ridiculous stuff,” he said.
