Fake news is too big and messy to solve with algorithms or editors — because the problem is….us.
Increasingly, I’m frustratedby (and often antagonistic toward) the emergent narrative about how to address so-called “fake news.” My anger is growing, not only because as I write this I’m almost 10 months pregnant and grouchy, but also because I see the possibility of well-intended interventions backfiring. I understand why folks want to do something now — there’s a lot of energy in this space, and the underlying issues at play have significant consequences for democracy and society. Yet what’s happening at this present moment is not actually new. It’s part of a long and complicated history, and it sheds light on a variety of social, economic, cultural, technological, and political dynamics that will not be addressed through simplistic solutions. Racing to implement Band-Aids may feel good, but I worry that such an approach creates a distraction while enabling the underlying issues to flourish.
Let’s start with a common “fix” that I’ve heard in the solutionist mindset: Force Facebook and Google to “solve” the problem by identifying “fake news” and preventing it from spreading. Though I appreciate the frustration over technology companies’ ability to mirror and magnify long-standing social dynamics, regulating or pressuring them to find a silver bullet solution isn’t going to work. From my vantage point, this approach immediately makes visible three differently scaled problems:
1) No one can even agree on a definition of “fake news,” even though a ridiculous number of words are being spent trying to define it.
2) Folks don’t seem to understand the evolving nature of the problem, the way that manipulation evolves, or how the approaches they propose can be misused by those with whom they fundamentally disagree.
3) No amount of “fixing” Facebook or Google will address the underlying factors shaping the culture and information wars in which America is currently enmeshed.
What is “Fake News?”
I’m not going to try to create a masterful definition of “fake news,” but I do want to highlight the interwoven tropes that are at play. Discursively, this frame is used to highlight every form of problematic content, including both blatantly and accidentally inaccurate information, salacious and fear-mongering headlines, hateful and incendiary rhetoric produced in blogs, and propaganda of all stripes (driven by both the State and other interests). Throughout my career, I’ve watched the deployment of such slippery terms (including bullying, online community, social networks, et cetera) for all sorts of political and economic purposes, and I have consistently found that without a precise definition or a clearly articulated problem, all that is achieved from drumming up conversations about the dangers of XYZ is spectacle.
By and large, I mostly see “fake news” being deployed by folks as a new frame for pushing long-standing agendas and commitments. This is true for researchers who have long critiqued corporate power, and it’s true for conservative pundits who love using this frame to shore up their hatred for mainstream media. So now there are dozens of meetings being held on “fake news” as folks wring their hands to find a solution; meanwhile, pundits and advocates of all stripes are calling on companies to fix the problem without even trying to define the problem. We’re seeing some folks focusing intently on “accuracy” and “truth,” while others are more focused on how content produces cultural frames.
Inside the internet platform companies, meanwhile, folks are struggling to create content policies that can be consistently applied. I’m always astonished at how inconsistent people are about what should and shouldn’t be prevented under the umbrella of “fake news” — and I’m mostly talking to experts.
Opening up the process doesn’t help. When the public is asked to report “fake news,” reports stream in from men’s rights advocates calling feminist blog posts critiquing patriarchy “fake.” Teenagers and trolls challenge pretty much everything.
Finding a third party to turn to isn’t much better. Experts can’t even agree on who is a hate group and who is engaging in protected speech. (I’m partial to the SPLC list, but that shows my political bent. And even its list doesn’t account for all of the groups that progressives might want to label as hate groups.) Just ask folks where they stand on blocking tabloid journalism or Breitbart and you’ll see conflict immediately.
Although a lot of the emphasis in the “fake news” discussion focuses on content that is widely spread and downright insane, much of the most insidious content out there isn’t in your face. It’s not spread widely, and certainly not by people who are forwarding it to object. It’s subtle content that is factually accurate, biased in presentation and framing, and encouraging folks to make dangerous conclusions that are not explicitly spelled out in the content itself. That’s the beauty of provocative speech: It makes people think not simply by shoving an idea down their throats, but inviting them to connect the dots. That content is far more powerful than content that suggests that UFOs have landed in Arizona.
As we distract ourselves with content produced for economic gain (that is often forwarded more by those who are aghast by the sheer inaccuracy of it than those who believe it), countless actors are building the capacity to manipulate others through much less easily detectable content. They’re hacking the attention economy and they keep iterating as people try to block different things that they try. This is why “meme magic” is so impressive — it’s about setting in motion particular frames and logics that can be activated through memetic reference, making it harder and harder to stop.
Increasingly, problematic content is shifting from being text-based to visual based, making it even harder to understand and address because of how it relies on all sorts of cultural references and symbols. It can be read by some as ironic or critique, while others see it as reifying and validating of ideas. Just do a search for how the swastika is used in cartoons for all sorts of political commentary. How you read those images has a lot to do with how you perceive the swastika.
To my frustration, even as pressure on companies to do something — anything — has increased, I have yet to see a robust proposal for which content should be removed through what process. It’s all a matter of “they” should just do it. Don’t get me wrong — there are some great low-hanging fruit mechanisms to cut off economic sources (although Google killing off AdSense for some sites has prompted other ad networks to step in). And I’m in favor of proposals that try to combat clickbait-style forwarding without reading — this forces folks to do more work before spreading something based purely on a headline. But at the end of the day, these are rounding errors in the ecosystem, even though folks seem to think that they’re big wins.
A decade ago, when I was an “ethnographic engineer” at Blogger, I spent a lot of time wading through customer service complaints, randomly sampling blog posts and comments, and building small tools to understand the nascent blogosphere and address problematic content. Like my earlier work on Usenet and my subsequent mapping of Twitter practices, I was astounded by the sheer ingenuity of people who were able to manipulate any well-designed feature that we implemented—see, for example, the rise of pro-ana in response to efforts to block anorexic content. In short, when AOL and other services blocked references to “anorexia,” those who identified anorexia as a lifestyle started referring cryptically to their friend “Ana” as a coded way of talking about anorexia without trigger the censors. Efforts to block certain language often trigger innovative ways around it.
These dynamics are age-old. Many who think about technology recognize the battle that emerged between spammers and companies over email, and the various interventions that emerged to try to combat spam. (Disturbingly for most decentralization advocates, centralization of email by Google was probably more effective than almost any other intervention.) Search engine optimization is another battleground that continues to plague the ecosystem. (Disturbingly for most anti-surveillance folks, this one has been curbed mostly through “personalization” algorithms that make mass targeting less effective.)
Part of why folks are targeting Google and Facebook in the “fake news” debate right now is that they have an effective monopoly on online information flows in certain segments of society. Because centralized systems have been able to somewhat curb spam and SEO, it seems that they should be able to stop this—except that folks really don’t like the “personalization” solution that emerged in 2016 when folks previously complained about problematic content.
Unfortunately, the problem space we’re talking about with “fake news” is way different than the problem space in spam or SEO. For one, the motivations behind the various folks seeking to manipulate content are extremely diffuse. If it were just about money, we’d be having a different conversation. But even in the money-centric space, look at how contemporary product marketing works.
Even if the goal were to curb the most egregious lies for economic gain (or even just deception in business in FTC parlance), that conversation wouldn’t be quick or easy — folks forget that iterations in spam/SEO went on for decade(s) to get to the current status quo (which is still imperfect but less visible, especially to Gmail users and sophisticated American searchers). These are international issues with no good regulatory process or reasonable process to adjudicate what is real or not. Welcome to the high stakes game of whack-a-mole.
Try writing a content policy that you think would work. And then think about all of the ways in which you’d be eliminating acceptable practices through that policy. Next, consider how your adversaries would work around your policy. This was what I did at Blogger and LiveJournal, and I can’t even begin to express how challenging it was to do that work. I can’t even begin to tell you the number of images I saw that challenged the line between pornography and breastfeeding. These lines aren’t as clean as you’d think.
I don’t want to let companies off the hook, because they do have a responsibility in this ecosystem. But they’re not going to produce the silver bullet that they’re being asked to produce. And I think that most critics of these companies are really naive if they think that this is an easy problem for them to fix.
Going Beyond the Surface
Too many people seem to think that you can build a robust program to cleanly define who and what is problematic, implement it, and then presto—problem solved. Yet anyone who has combatted hate and intolerance knows that Band-Aid solutions don’t work. They may make things invisible for a while, but hate will continue to breed unless you address the issues at the source. We need everyone — including companies — to be focused on grappling with the underlying dynamics that are mirrored and magnified by technology.
There’s been a lot of smart writing, both by scholars and journalists, about the intersection of intolerance and fear, inequality, instability, et cetera. The short version of it all is that we have a cultural problem, one that is shaped by disconnects in values, relationships, and social fabric. Our media, our tools, and our politics are being leveraged to help breed polarization by countless actors who can leverage these systems for personal, economic, and ideological gain. Sometimes, it’s for the lulz. Sometimes, the goals are much more disturbing.
Although most techies imagined that their tools would be used to bridge disconnects, that did not happen. But it’s not just tech. The ideals of journalism aren’t manifested in the current instantiation. Hell, even the ideals of market-oriented capitalism aren’t actually manifested in the corrupted instantiation that we currently have where finance can manipulate business (and everything else) for greedy purposes.
Part of the heady challenge in all of this is that we’re all trapped up in a larger system that’s deeply flawed. Sure, there are individual actors who are especially greedy or malicious in their intentions, but the banality of evil runs deep. So how do we channel our collective anger, frustration, and energy in a direction that goes beyond Band-Aid solutionism? How do we push past the desire to make the disconnect less visible and work to rebuild social infrastructure and bridge social divides? Above all, how do we not feed into the hands of polarization?
In my head, the design imperative that we need to prioritize is clear: Develop social, technical, economic, and political structures that allow people to understand, appreciate, and bridge different viewpoints. Too much technology and media was architected with the idea that just making information available would do that cultural work. We now know that this is not what happened. So let’s put that goal at the center of our advocacy and development processes and see what we can build if we make that our priority. Imagine if VCs and funders demanded products and interventions that were designed to bridge social divide. How can we think beyond the immediate and build the social infrastructure for the future? I’m not sure that we have the will, but I think that’s part of the problem.
The puzzles made visible through “fake news” are hard. They are socially and culturally hard. They force us to contend with how people construct knowledge and ideas, communicate with others and construct a society. They are also deeply messy, revealing divisions and fractures in beliefs and attitudes. And that means that they are not technically easy to build or implement. If we want technical solutions to complex socio-technical issues, we can’t simply throw it over the wall and tell companies to fix the broken parts of society that they made visible and helped magnify. We need to work together and build coalitions of groups who do not share the same political and social ideals to address the issues that we can all agree are broken. Otherwise, all we’re going to be doing is trying to wage a cultural war with companies as the intermediary and referee. And that sounds like a dreadful idea.