As major tech companies struggle with their responsibilities to the users and communities they serve, the question of what political candidates should be allowed to say in their online postings has emerged as a flash point. Each of the various platforms have taken different approaches to this question, but they each ignore the most important consideration of all: Why is it more advantageous online for a political candidate to be sensational rather than measured?
One in five American adults report getting their news from social media, and they’re spending more than six hours every day online. The internet, most of which is controlled by a handful of large corporations, has become our public square. Political candidates and their campaigns will always go where their voters are.
So far, tech companies’ first instinct with respect to political speech by candidates has been censorship (literally, the alteration or removal of content), which includes banning, restricting features, limiting reach, and labeling posts with warnings. Censoring a candidate’s political speech—whether it is done by a government or private company—is never a neutral act. A value judgment is always involved, and it is unlikely that we would ever reach consensus on the degree of censorship that is appropriate.
In recent weeks, we’ve witnessed two divergent approaches from Facebook and Twitter in response to identical language from President Trump. In one instance, Twitter added a label to a tweet which it claims violated its policies, limited its distribution, and disabled retweeting and other functionality. In a second instance, Twitter hid a tweet from immediate view and similarly curtailed people’s ability to share and comment on it. Facebook, in contrast, took no action in either case and critics have vocally opposed the company’s decision.
Instead of expecting Big Tech to police the content of politicians’ posts, we should pressure these owners of our public square to remove the incentives for candidates to post content designed solely to stimulate an algorithm. Hyperbole and emotion are rhetorical devices as ancient as the practice of politics, but only with the advent of products from Facebook, Twitter, and Google has there been a clear strategic incentive for candidates and their campaigns to operate exclusively at or beyond the extremes of political discourse.
Hyperbole and emotion are rhetorical devices as ancient as the practice of politics.
While the companies won’t disclose exactly how their algorithms work, we do know that engagement is one of the most important signals their platforms rely upon to judge how widely to distribute content. The more reactions, comments, and shares a post on Facebook receives, the more users it will reach. The same is true for Twitter, where, due to its decision to ban all political advertising, the only way for candidates to reach more voters is through likes and retweets. The more extreme a candidate’s message is, the more likely it is to provoke emotional reactions, both positive and negative, which results in more engagement and reach.
The algorithmic influence isn’t restricted just to social media. The percentage of supporters who open and click on links in an email determines whether a campaign’s future messages reach the precious real estate of the inbox or get banished to the spam folder. Google Search ads that generate more clicks are deemed more relevant to a user’s query and cost the advertiser less. On YouTube, the more time users spend watching a video—not its quality—determines how often it is recommended to others.
These effects may seem small, but in the world of campaigns, they add up to make a big difference. An analysis of Facebook ad spending by the Clinton and Trump campaigns in 2016 reveals that Republicans’ ability to understand the site’s mechanics gave them an advantage in both reach and cost.
In 2020, as Big Tech addresses the unique challenges presented by political speech, democracies worldwide are reckoning with the fact that the internet has driven political campaigns to become high-performance marketing operations.
In the current attention economy, a candidate’s opponent isn’t simply the other party’s nominee, but all of the other things that a voter could do with his or her time. More than a battle between Republican and Democrat, campaigns are competitions between democracy and apathy. Your elected officials spend every day in a pitched battle vying for your attention, which is most easily achieved by being as extreme as possible.
With the stakes of our elections so high, do we really think the winners should be determined by which campaign is the best at exploiting algorithms?
Do we really think the winners should be determined by which campaign is the best at exploiting algorithms?
You may ask yourself, shouldn’t we expect better of our politicians? But the fact is campaigns are laser focused on four key objectives: winning votes, getting attention, raising money, and upsetting the opposition. Anything outside of that matrix is a distraction.
Fortunately, we have existing models in the analog world for dealing with the complexities of political advertisements by candidates. In many offline cases, political speech made by officially filed candidates for public office is treated differently than other entities on the same platform. These two requirements, a formally registered candidate and political speech, are at the essence of this approach.
Under federal law, broadcast TV stations must sell ad time to legally qualified candidates at the “lowest unit rate” and their opponent must also be able to buy similar time at the same rate. Most importantly, broadcast TV stations are prohibited from censoring a candidate’s ad for any reason and are shielded from potential liability as a result.
The U.S. Postal Service provides candidates and other political entities reduced rates, and regulations state that “any political campaign mailing must be expedited through postal operations.”
So how might such a system work with major tech platforms?
Facebook and Twitter could guarantee a minimum level of distribution—regardless of engagement—for a certain number of organic posts and paid ads by political candidates for a length of time leading up to an election, thus removing the strongest incentives for emotionally charged rhetoric. The candidates would still be in control of their posts, but knowing that most of their constituents will see a post removes the need to rely on going viral to reach the broadest possible audience.
The platforms could even hide like, reaction, share, and retweet counts from these posts to further bolster the effect.
Similarly, email inbox providers like Google, AOL, Yahoo, and Comcast could provide candidates with a DNS record to ensure that candidate emails will reach inboxes.
The best news of all is that this solution doesn’t require government intervention.
Rather than competing in auction-based online advertising pricing, Facebook, Google, and other ad networks could set fixed prices and guaranteed availability for candidate advertising to eliminate any advantage that might be won with a more incendiary ad.
The best news of all is that this solution doesn’t require government intervention and could be taken up by tech companies on their own if they find the will. At this very moment, the tech giants are under enormous pressure from both within and without. Adopting this recommendation could represent significant progress in strengthening our democracy.
This solution is also adapted from existing infrastructure and could be deployed in time for this year’s elections. Both Facebook and Twitter already have systems in place for verifying and designating candidates for office. With email, protocols already exist to verify the identity of a sender to prevent users from domain spoofing.
While the technology Google, Facebook, Twitter, and others have created is new, the problems they face related to politics are not. Ironically, broadcast TV and postal mail might be the sources of disruptive innovation Big Tech needs to start navigating this challenge. However, extreme political rhetoric may continue to flourish unless tech platforms decide to reduce their algorithms’ focus on engagement and sensationalism for good.
Will this make campaigns more civil? There’s no guarantee, as it will always be in a politician’s nature to push rhetorical limits. But it will help move us past this fraught debate over censorship and start shifting elections away from algorithmic battle.