Like many of you, I’ve spent a lot of time during the past few months frustrated by the spread of disinformation on social media. As 2020 crawls to a conclusion at the speed of a Georgia vote count, I thought it might be useful to ask an expert to explain what passes for rules on the major platforms. Fortunately, we have just the right person on staff at ProPublica’s virtual headquarters:

I’m Kengo Tsutsumi, ProPublica’s platform editor and a member of our audience team. I spend my time thinking about how best to get our journalism in front of people and where we should focus to make that happen. There are lots of ways to do this, and social media is one of them. I’m as exhausted as the rest of you after the past few weeks, but to help make things less confusing, I thought it might be helpful to share some things I know to be true about social platforms.

It might seem like social media is the Wild West, with hardly any sheriffs and a set of rules that seemingly change from day to day. That’s not quite true, though it certainly has looked like that in the past four years. In fact, the biggest platforms — Twitter, Facebook, Instagram — do have rules which are enforced ... extremely selectively. And therein lies the problem.

Twitter, for example, has “terms of use” that, among other things, bar people from threatening violence or “hateful conduct.” For example, in November, Steve Bannon was permanently banned from Twitter (but not Facebook!) for calling for Anthony Fauci’s beheading on his podcast. Rapper Talib Kweli’s account was likewise permanently suspended over the summer for targeted harassment. In October, Charlie Kirk, a conservative activist, was temporarily suspended from Twitter for spreading false voter information when he shared a ProPublica and Philadelphia Inquirer story, incorrectly stating that it showed Pennsylvania had rejected 372,000 mail-in ballots. (We reported that 372,000 ballot requests, most of which were duplicates, were rejected, not ballots themselves.)

Up to this point, it all probably sounds pretty logical. Ignoring the fact that lots of tweets flagged for breaking the rules stay posted, the system really collapses when someone with, say, 85 million followers who happens to be president of the United States starts tweeting outright lies. By any of Twitter’s definitions (even under the public service exceptions Twitter outlines for world leaders), President Donald Trump could have been kicked off the platform many months ago, for example when he called for shooting Black Lives Matter protesters (the tweet was flagged but remained on the platform) or when he explicitly threatened other countries and their leaders or any of the number of times he’s insulted or attacked individuals.

There was clearly little stomach for removing the nation’s leader from social media. Instead, by the 2020 election both Facebook and Twitter chose to label presidential posts that spread clearly untrue election information as “disputed.” At times, the platforms appended factual information completely at odds with the president’s version of reality. That left us with a number of posts that looked like this:

You might be asking yourself, Why can Twitter and Facebook just make up rules and enforce them inconsistently? First, they are companies and courts have given corporations broad latitude to set rules of behavior on their own property, which in this case is their website. The basic parameters were spelled out in the foundational statute of the modern internet: the Communications Decency Act of 1996, specifically Section 230. The internet was in its infancy back then; Mark Zuckerberg was 12 years old and wouldn’t write the first code for Facebook for another six years. Much of the web involved message boards on which people posted whatever came to mind, including the sorts of lies that can get a newspaper or magazine sued for libel.

Entire books have been written about Section 230, but essentially what it says is that the provider of an interactive computer service (e.g., Facebook and Twitter) is not responsible for what users of the service post. As the owner of the service, though, the company is allowed to create and enforce certain standards. This protects the owners of the service from being responsible for content that breaks the law while allowing them the freedom to suppress or remove content that they do not want to host for a wide variety of reasons. In other words, the platforms have a lot of leeway.

Similarly, the platforms are entirely free to make rules that suppress or boost stories written by media organizations. If they decide the best way to make money is to show you fewer news stories and more political posts from your friends and relatives, they’re entirely free to do so. A few years ago, that’s exactly what they did.

Independent research shows that Facebook does not suppress conservative content, but complaints about “censorship” — which I think you’ll continue to hear in the coming years — ignore the crucial point that these companies make decisions about what you see all the time with little transparency about what or why.

Earlier this year, the computer algorithm that powers Facebook decided that ProPublica was posting clickbait, stories with flashy headlines that promise salacious or highly newsworthy content but don’t deliver either. Our Facebook account was suppressed for weeks while the reps we work with at Facebook tried to find a human who could help. Ultimately, Facebook could never explain the source of the problem to us and we had no recourse. \

Over the past year, my colleagues on ProPublica’s audience team and I have pursued a strategy of using social media platforms to share the “receipts” that underlie our often-complex investigative stories. Sometimes these documents are actual receipts; sometimes they’re images, letters or videos. It’s my hope that you’ll continue to see that material on Facebook, Twitter and elsewhere in the coming months and years. Unfortunately, what you see is ultimately not up to people like me.

Have a great week,

Kengo