Shield us From Viral Algorithms!

Our daily lives may have changed, but Facebook hasn’t. If social media really wants to help us navigate this moment, they need to turn off the algorithms. 

This, dear reader, is the text of an Op Ed I wrote in early 2020. I sent it to several media outlets but was unable to find anyone who would publish it at the time. Was I right about the state of the world? Looking back now, you be the judge...

With the world in lockdown, we’re increasingly reaching for social media to connect to our workplaces and customers, families and friends. The very companies we recently detested for how they mishandled our information and our trust seem to be stepping up: Facebook is supporting small businesses, Google Classroom is teaching our children, Amazon is delivering our groceries, and Twitter’s CEO Is sending money to the frontlines.  It’s a dream come true for those very same companies who were only too recently reeling from “techlash.”

But while our lives offline have changed considerably, nothing has changed online. If anything, it’s getting worse. Misinformation about the pandemic is rampant. Americans are waking up to the fact that the patterns of information spread online combined with rampant misinformation are making things worse. To solve this problem, we need to bring the tools that fight viral spread offline to bear on this online threat.

We’ve all looked to the internet to find the information we need to defeat the coronavirus. But social media companies are attention machines, not information sources. Facebook, Twitter, and Google use algorithms to keep our eyeballs glued to their screens and keep their advertising revenue rolling in. These algorithms don’t care about human benefit, or societal harm, or even truth: just whether something is clicked, liked, or shared. Unfortunately, we are most likely to click, like or share when we feel outrage. Because of this, Google and Facebook lock us into filter bubbles and YouTube recommendations trend toward extremist content.

As such, the algorithms of attention, personalization, and search are now contributing to the viral spread of misinformation. A new study from Carnegie Mellon revealed that a lot of the 'groundswell’ of protest organizing against masking is astroturfed – delivered by fake accounts.  Twitter users repeat rumors that people with pure African blood are immune to Covid-19, that the virus is a liberal media hoax, that 5G network towers are to blame. No wonder Brits are burning cell phone stations and defiant Americans are hosting “coronavirus parties.” This is the price of platforms that are optimized for engagement.

Sociologists who study viral behavior show that hearing a story from just two distant sources is enough to make us spread that information ourselves. This kind of distant tie is common among online friends and followers. Unfortunately, we also know that people cannot tell humans from bots or detect impersonators of friendly neighborhood sources online. Hence Covid-19 offers perfect conditions for bad actors to slip fake stories into the stream, rife for algorithmic promotion to the top of your feed. Amid these sociological facts and coronavirus uncertainty, it is often impossible for well-meaning users to tell the difference between conspiracy and reality.

Search optimization algorithms are also to blame. Developed to help us sift through the vast array of content online, now they strategically direct our attention. If you, like everyone in the world, are looking for masks, hand sanitizer, or cures for the coronavirus, beware. Every snake oil salesman worth their salt is using Covid19 search terms to direct people to their wares.

When “preference,” “engagement,” and “attention” algorithms built for advertising are let loose on a global crisis, it has real implications. Online personalization makes it easy to get false information in front of vulnerable eyes in times of need. Some communities might not get the information they need to confront the virus precisely because of these tools.

Fortunately, fighting viral misinformation online looks a lot like fighting a virus offline.  We can start by improving our internet hygiene. Use systems that don’t optimize for your attention, sorting stories based on preference or outrage. Texting and making phone calls, or searching using DuckDuckGo are a good start. We can make separate online accounts for activities that require Google or Facebook, like our children’s activities or storefronts, and download browsers like Firefox that restrict Facebook tracking. And always trace information to its source before = sharing stories online, or purchase from untrustworthy sellers. Think of it as the online equivalent of handwashing.

There is an online equivalent to social distancing too. Avoiding these platforms entirely is not a credible option for those whose professional, parenting, and social lives have migrated online. But we don’t need distance from each other online: we need distance from algorithms.  Taking personalization algorithms out of service entirely will slow our transmission of misinformation, direct search results toward trustworthy sources, and reduce the effects of harmful viral conspiracies on vulnerable communities, while keeping us connected to those we love the most. 

That’s why companies like Google, Facebook, Amazon, and others must turn off the algorithms that track, promote, and amplify our connections online -- immediately. While even a few months ago these algorithms were the source of wild profits, unplugging them is now is actually in these companies’ best interest. After all, our concerns for masks and hand sanitizer won’t last forever or tell us apart for advertisers, and fake products and fake friends only decrease our trust in their services.

To be sure, this will pull the plug on the personal data economy: a thriving sector of personalization, tracking, and data analysis. It may seem easier to l play misinformation whack-a-mole with armies of content moderators working around the clock, as Facebook is currently doing, than it is to dismantle an entire consumer surveillance system. But that solution only addressed the symptoms, not the cause. And they do not help us to avoid the next onslaught. After all, COVID fake news practices are the perfect dry-run for influence peddlers looking to distort American democracy in November.

There is some good news. WhatsApp recently changed their system so that users can’t forward a message more than 5 times, or copy it from one group chat to another. This change is makig for an 80% reduction in Fake news circulation. It shows that making simple changes to the software behind the screen can make an impact in reducing the flow of misinformation.

The rampant spread of misinformation online during Covid-19 reveals that the cost of our favorite free services is too high. Algorithmic distancing and improving our internet hygiene will help us to save lives while keeping us socially connected. It’s not just up to us: internet companies can do their part by turning off their algorithms of personalization, engagement, and amplification. Flattening the curve offline requires working together online too.