BLOG : POST

Pay Attention to the Weak Links

Photo by amontel via Flickr (CC BY-NC 2.0)

Over the past several days, I’ve kept seeing this explanation of a scholarly paper on the epidemiology of disinformation. (Thanks to Laura Hazard Owen at Nieman Lab for first putting the story on my radar, and to my New School colleague David Carroll for flagging it again.)

Three scholars, who Bloomberg View contributor Mark Buchanan describes as “network theorists,” used deep reinforcement learning (the same approach to artificial intelligence that humbled Go master Lee Sedol) to simulate how fake news moves through social networks.

Buchanan writes that precise audience targeting was “the most important catalyst“ for false information to spread:

The key was to seed an initial cluster of believers, who would share or comment on the item, recommending it to others through Twitter or Facebook. False stories spread farther when they were initially aimed at poorly informed people who had a hard time telling if a claim was true or false.

In other words, purveyors of false information are looking for weaknesses in our information ecosystem so they can spread their lies.

This insight reminded me of an episode of Malcolm Gladwell’s podcast, Revisionist History, in which he distinguishes between “strong-link” and “weak-link” games. Gladwell characterizes basketball as a strong-link game, since teams invest heavily in their Big Three (LeBron, Kyrie, K-Love; Duncan, Parker, Ginobili; Bird, McHale, Parish) and fill out the bench with journeymen and rookies. In soccer, meanwhile, the least skilled player on the field can sink your team, so owners have spend more evenly among their top 11. 1

On the episode, Gladwell talks about $100 million donations to universities, which keep going to the Stanfords and Harvards instead of places like Glassboro State College, now Rowan University (and the exception to the megadonor rule). Gladwell argues that higher education in America is being played like a strong-link game, but in fact too many resources are devoted to a handful of elite institutions cultivating genius, while not enough is going to hundreds of others turning out competent engineers (or liberal arts majors).

On top of that, elite schools get disproportionate press attention as well. In his FiveThirtyEight story Shut Up About Harvard, Ben Casselman argued that reporters and editors cover these places so much because “at least in major national media outlets, that’s where most of them went.”

via GIPHY

Journalism has a similar problem. (The hiring patterns too, but let’s talk about the money first.) Digital media has concentrated investment and jobs in New York, Washington, and Los Angeles while local newspapers around the country are withering. As for nonprofit efforts, NYU’s Rodney Benson told Nieman Lab:

I think what a lot of foundation-supported media are doing is providing quality news to audiences that are already getting a lot of quality news. That’s not a bad thing, but I don’t think they’re addressing the problem of the broader lack of public knowledge in the larger citizenry.

In 1972, when Woodward and Bernstein started digging into Watergate, the United States had three national TV networks, which were largely fed by a few dominant national newspapers. The strongest channels for bogus stories back then were the supermarket checkout line and drive-time radio.

Newsrooms today can’t control how information is delivered, and must compete on the “even” playing fields of Google search, Facebook sharing, and Twitter, where sources and subjects speak directly to the public, including the president. It’s not only because of digital technology; direct mail and talk radio came before 4chan and Breitbart. But the internet has finally exposed this change to a mass audience, and even to executives and eggheads who don’t do their own grocery shopping.

The New York Times, the Washington Post, Buzzfeed, and ProPublica have been churning out great scoops about incompetence and corruption in the Trump Administration, but those stories haven’t yet had the impact that journalists imagine or readers hope. I can’t count how many nights this year I’ve gone to bed with Twitter-fueled fantasies that some story has dealt a decisive blow to some official or policy, only to wake up to the next morning to find that nothing has changed.

It’s easy to forget that it took Woodward and Bernstein time to build a case, and they did not act alone. They were a part of a larger information ecosystem, including Ben Bradlee and other editors, Deep Throat and other sources, rivals at other papers, wire services, and the three networks, and even the CBS employee (or was it a contractor?) who designed the Watergate icon for the evening news, which I learned to recognize as a toddler, since my parents watched Walter Cronkite every night.

Today’s information ecosystem is more complex, and while platforms like Facebook, Google, and Twitter have more power than the major papers or the TV networks had, I’m not convinced that Mark Zuckerberg could fix everything if he wanted to.

There may be some short-term solutions; in their paper “Fake News in Social Networks,“ Christoph Aymanns, Jakob Foerster, and Co-Pierre Georg say that “exposing agents to the possibility of fake news” (and these agents represent readers) proved to be an effect defense mechanism. Another is projecting personal privacy so it’s harder to figure out who to target.

But when you patch one vulnerability in a system, someone will find another to exploit. The price of liberty is eternal vigilance.

In any case, 2017 does not feel like the right time to focus on which news organizations poached which reporters from a rival, or for for-profit newspapers to seek philanthropic funding.

As that FiveThirtyEight headline writer would say, Shut Up About the New York Times. The weak links desperately need our attention. And weak links pose a threat to us all.

via GIPHY

  1. Basketball and soccer fans may quibble with this Gladwellian dichotomy, and I have dismissed others, such as his distinction between rock and country songs.↩︎