On September 8, 2004, longtime CBS anchor Dan Rather presented four documents about then-President George W. Bush’s Texas Air National Guard service in the early 1970s. Bush’s service had been a matter of some controversy throughout the 2004 presidential campaign, particularly given the rather suspect way in which his military records were released: on multiple occasions records deemed to have been lost were mysteriously found weeks or months later, and only the Bush campaign had managed to produce records placing the President in Alabama in 1972, when the Democratic National Committee accused him of being AWOL.
Rather was certain he had a smoking gun: documents from Bush’s squadron commander which detailed top-down pressure to adjust Bush’s record to appear more favorable than it appeared. The problem for Rather and CBS, though, is that the then-thriving blogosphere quickly demonstrated that the memos were almost certainly created in Microsoft Word with the default settings; the source for the memos, meanwhile, claimed to have burned them after faxing them to CBS, which, contra Rather’s claims, had not authenticated the documents.
CBS took the mistake seriously: the network commissioned a 234 page report by former U.S. Attorney General Dick Thornburgh and former Associated Press CEO Louis D. Boccardi that concluded:
The stated goal of CBS News is to have a reputation for journalism of the highest quality and unimpeachable integrity. To meet this objective, CBS News expects its personnel to adhere to published internal Standards based on two core principles: accuracy and fairness. The Panel finds that both the September 8 Segment itself and the statements and news reports by CBS News that followed the Segment failed to meet either of these core principles…
While the focus of the Panel’s investigation at the outset was on the Killian documents, the investigation quickly identified considerable and fundamental deficiencies relating to the reporting and production of the September 8 Segment and the statements and news reports during the Aftermath. These problems were caused primarily by a myopic zeal to be the first news organization to broadcast what was believed to be a new story about President Bush’s TexANG service, and the rigid and blind defense of the Segment after it aired despite numerous indications of its shortcomings.
CBS responded to the report by apologizing to viewers, firing the producer who obtained the forged documents, while Rather retired (it is widely assumed but not confirmed that Rather’s retirement was because of the controversy).
What is notable about this episode is that it was in many respects the pinnacle of how the Internet could make traditional media better: CBS got duped, both because it wanted to be first and also, one suspects, because of confirmation bias (Rather continued to argue that the story was true even if the documents were false), but in this case, there were many more outlets than the big three news networks, and those outlets, in a classic example of “more speech” leading to the truth, corrected the misinformation. And CBS, to its credit, corrected the record.
Of course bloggers are particularly well-suited to identifying the finer points of Times New Roman; they are not so strong at international reporting, particularly when it comes to a regime like Saddam Hussein’s. And, in the case of Judith Miller’s reporting about Iraq’s alleged weapons of mass destruction in the New York Times, this bit of misinformation worked to Bush’s benefit. As I wrote in 2016’s Fake News, this was particularly damaging:
Looking back, it’s impossible to say with certainty what role Miller’s stories played in the U.S.’s ill-fated decision to invade Iraq in 2003; the same sources feeding Miller were well-connected with the George W. Bush administration’s foreign policy team. Still, it meant something to have the New York Times backing them up, particularly for Democrats who may have been inclined to push back against Bush more aggressively. After all, the New York Times was not some fly-by-night operation, it was the preeminent newspaper in the country, and one generally thought to lean towards the left. Miller’s stories had a certain resonance by virtue of where they were published.
After Miller’s stories were shown to be bogus, New York Times editors wrote, in an unsigned editorial:
We have found a number of instances of coverage that was not as rigorous as it should have been. In some cases, information that was controversial then, and seems questionable now, was insufficiently qualified or allowed to stand unchallenged. Looking back, we wish we had been more aggressive in re-examining the claims as new evidence emerged — or failed to emerge…Complicating matters for journalists, the accounts of these exiles were often eagerly confirmed by United States officials convinced of the need to intervene in Iraq. Administration officials now acknowledge that they sometimes fell for misinformation from these exile sources. So did many news organizations — in particular, this one.
Miller would leave the paper a year later, and while the New York Times didn’t quite live up to CBS’s standard of accountability, at least the Editors were honest that they had screwed up.
What Happened in 2016?
On October 6, 2020, Greg Bensinger, a member of the New York Times editorial board, exhorted people to Take a Social Media Break Until You’ve Voted:
Social media is a cesspool, and it’s getting worse by the day. In the past few months, outright lies about miracle coronavirus cures, mail-in voting fraud and Senator Kamala Harris’s eligibility for the vice presidency have gone viral on Facebook, Twitter, YouTube and elsewhere. People believe them. The platforms, however, aren’t taking the threat of spreading misinformation seriously enough ahead of the election. Again. That’s why I urge Americans to take a bold step: Stay off social media at least until you’ve voted.
This fits the New York Times’ preferred narrative of the 2016 election, in which social media generally, and Facebook specifically, was to blame for President Trump’s victory; the editorial board wrote in 2017:
Chastened by criticism that Facebook had turned a blind eye to Russia’s manipulation of the social network to interfere in the 2016 election, the company’s executives now acknowledge a need to do better and have promised to be more transparent about who is paying for political ads. That’s a good start, but more is required — of Facebook, of social media giants generally and of Congress.
Missing from the list was the New York Times itself, which as the Columbia Journalism Review argued in 2017, led the way in making the 2016 election about anything but the issues:
In light of the stark policy choices facing voters in the 2016 election, it seems incredible that only five out of 150 front-page articles that The New York Times ran over the last, most critical months of the election, attempted to compare the candidate’s policies, while only 10 described the policies of either candidate in any detail.
In this context, 10 is an interesting figure because it is also the number of front-page stories the Times ran on the Hillary Clinton email scandal in just six days, from October 29 (the day after FBI Director James Comey announced his decision to reopen his investigation of possible wrongdoing by Clinton) through November 3, just five days before the election. When compared with the Times’s overall coverage of the campaign, the intensity of focus on this one issue is extraordinary. To reiterate, in just six days, The New York Times ran as many cover stories about Hillary Clinton’s emails as they did about all policy issues combined in the 69 days leading up to the election (and that does not include the three additional articles on October 18, and November 6 and 7, or the two articles on the emails taken from John Podesta). This intense focus on the email scandal cannot be written off as inconsequential: The Comey incident and its subsequent impact on Clinton’s approval rating among undecided voters could very well have tipped the election.
Nate Silver said that is exactly what happened:
Hillary Clinton would probably be president if FBI Director James Comey had not sent a letter to Congress on Oct. 28. The letter, which said the FBI had “learned of the existence of emails that appear to be pertinent to the investigation” into the private email server that Clinton used as secretary of state, upended the news cycle and soon halved Clinton’s lead in the polls, imperiling her position in the Electoral College…
And yet, from almost the moment that Trump won the White House, many mainstream journalists have been in denial about the impact of Comey’s letter. The article that led The New York Times’s website the morning after the election did not mention Comey or “FBI” even once — a bizarre development considering the dramatic headlines that the Times had given to the letter while the campaign was underway…The motivation for this seems fairly clear: If Comey’s letter altered the outcome of the election, the media may have some responsibility for the result. The story dominated news coverage for the better part of a week, drowning out other headlines, whether they were negative for Clinton (such as the news about impending Obamacare premium hikes) or problematic for Trump (such as his alleged ties to Russia). And yet, the story didn’t have a punchline: Two days before the election, Comey disclosed that the emails hadn’t turned up anything new.
The lack of a punchline applies to many of the Facebook controversies since then: the United Kingdom’s Information Commissioner’s Office determined that the only scandal about Cambridge Analytica was the degree to which they oversold their capabilities;1 the afore-linked report from the Columbia Journalism Review highlighted how infinitesimal the scale of Russian interference on the platform was, and research shows that “fake news” makes up a fraction of American’s media diet; more recent research about voting fraud argued:
Contrary to the focus of most contemporary work on disinformation, our findings suggest that this highly effective disinformation campaign, with potentially profound effects for both participation in and the legitimacy of the 2020 election, was an elite-driven, mass-media led process. Social media played only a secondary and supportive role.
Just like her emails.
Twitter vs. the New York Post
It may seem odd to be re-litigating the 2016 election two weeks before the 2020 one, but last week’s decision by Facebook and Twitter to slow and ban respectively a sketchy story about Vice-President Joe Biden’s son Hunter cannot be understood without looking back to 2016. The Verge reported on Thursday:
Facebook has reduced the reach of a New York Post story that makes disputed claims about Vice President Joe Biden’s son, Hunter, pending a fact-check review. “While I will intentionally not link to the New York Post, I want be clear that this story is eligible to be fact checked by Facebook’s third-party fact checking partners. In the meantime, we are reducing its distribution on our platform,” tweeted Facebook policy communications manager Andy Stone.
Twitter banned linking to the Post’s report, but it cited a different policy: the site’s rules against posting hacked material. “In line with our hacked materials Policy, as well as our approach to blocking URLs, we are taking action to block any links to or images of the material in question on Twitter,” a spokesperson told The Verge. Clicking existing links will direct users to a landing page that warns them it may violate Twitter guidelines.
Twitter’s actions became even more extreme over the next few days, including banning follow-up stories from the New York Post and locking the newspaper’s Twitter account, even as the company’s explanation for its actions continued to shift; eventually the company unblocked the article, claiming that the story was now widely spread on the Internet.
The story, to be clear, appears to be fabricated, and comically so. The reason I think that is because the story has been relentlessly investigated and criticized both by other media outlets and people on social media — kind of like Dan Rather’s George Bush story was (both the New York Times and New York Magazine reported that the author of the New York Post story refused to put his name on it because he was so skeptical of the story’s sourcing).
This, to be very clear, does not exculpate Twitter’s actions: quite the opposite in fact. After all, as Twitter itself acknowledged, banning the link did not stop the news of the story from spreading. If anything Twitter’s actions had the opposite effect: it made the story spread far more widely than it would have otherwise, now with the additional suspicion that the powers-that-be must want to hide something. What was overshadowed were all of the stories making the case that the story may have been fabricated.
Stories Versus STORIES
Here’s the thing about social media: simply facilitating the transmission of information doesn’t make a story into a STORY; look no further than the coronavirus. I wrote in Zero Trust Information in March:
It is hard to think of a better example than the last two months and the spread of COVID-19. From January on there has been extensive information about SARS-CoV-2 and COVID-19 shared on Twitter in particular, including supporting blog posts, and links to medical papers published at astounding speed, often in defiance of traditional media. In addition multiple experts including epidemiologists and public health officials have been offering up their opinions directly.
That post was published the morning of March 11, when COVID-19 was still being mostly ignored. That all changed 12 hours later when (1) NBA player Rudy Gobert tested positive, leading the NBA to suspend play, (2) Tom Hanks announced on Twitter that he had tested positive, and (3) President Trump announced that the U.S. was suspending travel from Europe. Suddenly the coronavirus was a STORY that everyone knew about.
What is important to note is that no new facts about the coronavirus had emerged in those 12 hours: what mattered is who was talking about them. The NBA is an institution, Hanks a beloved actor, and Trump the President. Each is capable of creating a STORY in a way you or I cannot.
This, more than any other media entity, has long applied to the New York Times. Maxwell McCombs and Donald Shaw, in their seminal paper The Agenda-Setting Function of Mass Media, documented how stories became STORIES; McCombs noted in a book about the same topic:
The general pattern found is that the agenda-setting influence of the New York Times was greater than that of the local newspaper, which, in turn, was greater than that of the national television news.
This effect has certainly been diminished by the rise of cable news and the Internet, but even then, outlets like Fox News are often defined by their opposition to the New York Times, which is itself a testament to the New York Times‘s influence. It also explains why the Clinton email story, much like the Judith Miller stories a decade earlier, were such a big deal: it wasn’t simply that the New York Times was writing about a particular story, but they were writing about a story that seemed favorable to Republicans and a problem for Democrats. Surely then it must be important and true?
The point of this is not to debate whether or not the email story was true, or Hunter Biden’s laptop story. Rather, it’s to establish that while social media publishes everything, from mountains of misinformation and conspiracy theories to critical information about an impending pandemic, making something matter requires more than manufacturing zero marginal cost content. The New York Times has that power by default, while Twitter and Facebook only has that power to the extent they do the opposite of what most expect from them (which is to act as a utility for the conveyance of information).
Responsibility and Accountability
Thus the look back to 2016: the reason why I believe that the New York Times was more responsible than Facebook for the election outcome is rooted in a belief that making stories matter is far more important and impactful than making up stories. Unfortunately, the way in which the generally accepted narrative about the election shifted to blaming Facebook led to a crisis of accountability.
The first way in which accountability can go wrong is in the lack of it. Unlike CBS with the Bush papers, or the New York Times with its Iraq reporting, there has been little if any self-acknowledgment of the role the media played in the 2016 election outcomes. That is self-evidently bad.
What is more subtle, though, is the problems that comes from mis-assigning accountability. Making social media the scapegoat for 2016, for example, not only meant that there was no accountability for media coverage, but also led directly to Twitter wielding power it never should have.
It is, to be clear, absolutely outrageous that a communications platform unilaterally decided what was or was not true, outright barring a major publication from accessing its services, even if the story was false. Twitter’s role with regards to the Hunter Biden story should have been to facilitate more information sharing, in this case to disprove the story, not to arbitrarily decide what was or was not true. No, this wasn’t a crime — last week was also a demonstration that no platform has a monopoly on the distribution of information — but it was majorly at odds with the fundamental principles that undergird a liberal democracy.
It was also monumentally stupid: I noted this summer that Republican suspicion of tech was rooted not in traditional economic arguments about antitrust, but rather in political concerns about a traditionally Democratic industry controlling communications. Twitter basically wrote the case study for why, if you are concerned about the political problems entailed by major platforms, tech needs more regulation.
At the same time, I can understand why Twitter acted; Casey Newton applauded the move on Platformer, writing:
In the run-up to the 2020 election, platforms have been preparing for all manner of threats. One that they have warned about with some frequency is the “hack and leak” operation. A hack-and-leak occurs when a bad actor steals sensitive information, manipulates it, and releases it in an effort to influence public opinion. The most famous hack-and-leak is the dissemination of Hillary Clinton’s stolen emails in 2016, which may have affected the outcome of the election.
A hack-and-leak works because it exploits journalists’ natural fondness for writing about secret documents, ensuring that they get wide coverage — sometimes before reporters have a chance to closely examine their provenance. (It turns out that basically all humans have a fondness for reading secret documents, and one reason hack-and-leaks seem particularly threatening in the age of social networks is that platform sharing mechanisms allow these stories to spread around the world more or less instantly.)
Because of the role they play in amplifying big stories, platforms have taken the prospect of a hack-and-leak on the eve of the election quite seriously. And so when the New York Post dropped its story about a laptop of dubious origin containing what purported to be incriminating documents related to Joe Biden and his son, the Spidey senses of platform integrity teams all began to tingle in harmony.
Here is the issue: Newton rightly connected the dots to the Clinton email scandal, but instead of pointing the finger where it belonged — the journalists and publications that wrote about them incessantly — accountability was passed to the platforms, approvingly so:
In the run-up to the election, platforms have accepted two key responsibilities: to reduce the spread of harmful posts, and to reduce that spread quickly…A hack-and-leak operation represents one of the most difficult tests of this commitment — the operation is designed to spread far and wide long before all the real facts can be known.
There are cases in which I think platforms should act even faster than this — three hours is too long to decide whether to decide whether a single tweet from a high-profile account violates policy, I think. But to identify a potential hack-and-leak operation and restrict it within a few hours, before it hits the Twitter Trending page, deserves some credit.
Again, to be perfectly clear, Twitter’s actions made the story far bigger than it would have been otherwise. To applaud their action is akin to giving a participation trophy to a team that finished in last place: are we cheering the effort, or the actual results? As long as we care more about the former than the latter we are going to get more screw-ups from platforms looking to make up for mistakes that were never their responsibility in the first place.
This, though, is what I mean about the crisis of accountability. The reality is that tech has absolutely taken the criticisms about its role in 2016 to heart, and often to good effect: it is a good thing that Facebook is taking issues like foreign interference far more seriously than they did previously, and spending billions of dollars — to the point where it is seriously impacting its earnings — on security and moderation; Twitter too has gotten much more serious about policing its platform when it comes to everything from bots to abuse. Those are positives.
The problem is the temptation to go beyond the platform and start policing everyone else because you have been led to believe that you are accountable for everyone else’s failures. In fact, at the end of the day, it is the New York Post that is responsible for what it publishes, and it is a mistake for Twitter to take responsibility for that; the same applies to whatever is published by the New York Times, CBS, and any other media entity. Twitter (and Facebook) have an awesome amount of power, which means that both platforms needs to not only be more cognizant about what they should seek to control, but also about what they should not. And, by extension, that is why it was so damaging to shift responsibility in 2016, because you can understand why both platforms might feel compelled to control too much.
A Framework for Moderation
The framework I would suggest these platforms follow is based on one I developed last year in A Framework for Moderation:
It makes sense to think about these positions of the stack very differently: the top of the stack is about broadcasting — reaching as many people as possible — and while you may have the right to say anything you want, there is no right to be heard. Internet service providers, though, are about access — having the opportunity to speak or hear in the first place. In other words, the further down the stack, the more legality should be the sole criteria for moderation; the further up the more discretion and even responsibility there should be for content:
Note the implications for Facebook and YouTube in particular: their moderation decisions should not be viewed in the context of free speech, but rather as discretionary decisions made by managers seeking to attract the broadest customer base; the appropriate regulatory response, if one is appropriate, should be to push for more competition so that those dissatisfied with Facebook or Google’s moderation policies can go elsewhere.
The addition I would make to this framework is that responsibility accrues to the layer of agency, both as a matter of principle and of practicality. In other words, the New York Post is responsible for what they publish, and not only should Twitter not decide whether or not that is acceptable as a matter of principle, it needs to recognize that doing so will not kill the story but rather make Twitter’s abuse of power the story. To go in the other direction — to make every part of the stack responsible for everything in the stack — is to inevitably end up in a world of ISP-level control on what we can or cannot see.
This is why I think that Facebook’s decision to slow the spread of this story, as opposed to outright censorship, was a reasonable one; spread is what Facebook has agency over, not necessarily the story itself. This is also why Facebook’s banning of Qanon is acceptable in a way Twitter’s link ban was not: the former happens on Facebook, and is thus Facebook’s responsibility.
This also leads to two further points about conservative objections to Twitter’s actions specifically. First, removing Section 230 protections will not solve this problem, but rather make it worse: not only will Twitter and Facebook be more motivated to censor potential misinformation, but so will every level of the Internet stack.
At the same time, suspicion of tech’s power is clearly justified, and not only when it comes to Twitter and Facebook. Notice how Apple’s App Store and Google’s Play Store are increasingly levers for control, both by liberal democracies (as in the case of TikTok and WeChat) and authoritarian regimes (like Belarus or China). The former raise questions as to whether tech executives can be trusted with their power, and the latter questions as to whether they can resist it being utilized.
To that end, perhaps Twitter’s actions will ultimately prove to be a good thing, simply for the fact this power was so brazenly displayed on a story that was ultimately inconsequential. The best solution to too much power is to devolve it and increase competition, whether that be national anchors, national newspapers, or national water coolers, and it may be time to figure out what that devolution should look like.