the Dangerous
Rise of
"alternative" Facts

by Jonathan Dotan and Cheryl Phillips

Four years ago, one composite news image published by Reuters at 2:02 p.m. the day of the presidential inauguration was intended as a straightforward record of the day’s events.

Instead, it marks a moment when the role of photojournalists was reversed from reporting the news to becoming the news story itself. Reuters touched off a firestorm of social media activity when it published to its news wire the following twin photographs:

The caption read: “A combination of photos taken at the National Mall shows the crowds attending the inauguration ceremonies to swear in U.S. President Donald Trump at 12:01 p.m. (L) on January 20, 2017 and President Barack Obama on January 20, 2009, in Washington, DC.”

Those two images became the centerpiece of President Trump’s attacks on journalism about crowd sizes at the inauguration and a jumping-off place for other continued attacks on the press.

One reason Trump and others were able to cast doubt over how Reuters edited and presented the photos is that most photographs are published as just that —as images only. But now, four years later, Reuters and Starling Lab have collaborated to build a prototype of what photo authentication can look like, how it can push back against misinformation and disinformation, and how, going forward, it can support the important work of journalists.

Today, a photo’s visual pixels are just one part of a digital data set that can inform readers. It is possible to pair with each photograph the evidence that its contents have not been modified: stamps of the time, date, place of its capture; the servers where it has been stored; and finally, a transparent way to see how it has been reasonably adjusted by photo editors and validated by fact-checkers. All of this can be embedded directly into a photograph and travel with it no matter where —or on what platform— readers find it.

This archive of the 78 days between election and inauguration is not just a historical photo archive — it is a prototype. The reader can explore the embedded information that accompanies the photojournalism, from the moment a photo is taken to all the edits along the way.

As readers pore through the images tied to the 78 days of the transition, they can select a photo and click on the “i” icon in the corner, and see all of the information that went into the making of that image. But this is more than mere metadata. Each piece of information uses encryption to seal that stage in the process —to freeze it in time. An id or hash is created for each step in the process, putting a seal on an envelope of the data. Tamper with the seal, and it will be obvious. If the seal is intact, so is the underlying information.

On the other side of the 2020 election, reporters and editors continue to face unprecedented difficulties in doing even their basic job of providing the public with facts. This presidential election and transition is a clear sign of how much the media has been drowned out by misinformation and disinformation. The impact is evident. According to the Pew Research Center, more than half of U.S. adults have little to no confidence that journalists act in the public’s best interest.

Pause and consider this fact.

One in two people don’t just question the news media’s competence but believe that journalists are detrimental to society.

This alone is evidence of why we need journalists more than ever to re-establish a baseline of facts for democracy to function. This prototype by the Starling Lab and Reuters explores how a next-generation image archive can leverage advanced cryptography and decentralized web protocols to restore trust in images.

These tools may not stop attacks over fake news, but when assembled in our archive, they create a case study for what cryptographic authentication can mean going forward as we continue to innovate with digital publishing.

When questions over those inaugural photos by Reuters first took over social media in 2017, the news agency wrote about the process it followed to produce the images. It noted that it received special permission to access the observation level near the top of the Washington monument. A Reuters photographer hiked up 897 steps, carrying his camera, backpack, and related equipment, all so he could shoot a photo of Donald Trump’s inauguration from the same vantage point as President Barack Obama’s first inauguration.

But no amount of explanation stopped President Trump, who used his first speech in office to provide a different account of the events. “We had a massive field of people,” he proclaimed. “You saw them. Packed. I get up this morning, I turn on one of the networks, and they showed an empty field.”

“I have a running war with the media,” President Trump said. “They are among the most dishonest human beings on Earth.”

Later that same day, White House Press Secretary Sean Spicer used the administration’s first press briefing to continue the attack, suggesting images from Reuters and other media outlets were “intentionally framed” to minimize the crowds. He was flanked by monitors showing other photos from the day that supposedly made his case.

“This was the largest audience to ever witness an inauguration, period -- both in person and around the globe,” Spicer told reporters, exiting shortly afterward without accepting questions.

Setting aside both Trump and Spicer’s break with the minimum level of decorum typically afforded by the President to the press, the controversy over the photo provoked important conversations about how to establish facts from the published photographic work. In today’s world, any explanation of how a work is published is welcome, but we must not stop there. 

Photos can be hard to compare. Slightly different vantage points and cropping or editing an image can make a big difference in what is depicted.

The photos that Spicer selected to make his case show a view of a crowd that looks nearly full, and seemingly contradict the presentation of the Reuters photo. Another photo from CNN, taken with a gigapixel panorama camera, shows a different vantage point but a similar crowd size to what Spicer presented. This was not lost on various Trump supporters who made the case that the media was intentionally manipulating the images to diminish crowd sizes by comparing different photos.

In a swirl of different explanations, Trump’s supporters highlighted common problems in comparing photographs. They claimed the two photos were misleading because they were taken at different times. The photo of Trump’s inauguration, they argued, was taken earlier in the ceremony than the Obama inauguration photo. The smaller crowds were a result of attendees being significantly delayed from entering the grounds. Echoing grievances that persist to this day, commentators blamed Black Lives Matter and Antifa protesters for disrupting security checkpoints — a claim that exaggerated the scale of the protests near the park entrances and overstated their impact on spectator turnout.

However, both photos’ time frame was similar and multiple vantage points and even time-lapse video of the crowd from the entire event overwhelmingly support that Reuters’ had the clearest and most unobstructed view of the crowd.

Had those images been cryptographically sealed, perhaps many of those questions could have been answered by referencing the authenticated information accompanying the photos.

Instead, by the time the arguments aired, the photos were no longer a question of facts but politics. When pressed about the controversy, White House counselor Kellyanne Conway explained in a now-famous interview: “On this matter of crowd size, I think it is a symbol for the unfair and incomplete treatment that this president often receives.” Conway said. “You’re saying it’s a falsehood. … Sean Spicer, our press secretary, gave alternative facts to that.”

And as linguists and journalists agreed, “alternative facts” was a dangerous word to enter into the lexicon. Even Merriam-Webster thought it fit to remind us all the proper definition of facts.

Subsequent reporting uncovered the lengths to which the administration went to advance their false narrative. Trump personally intervened with the National Park Service to get the most favorable photographs, The Guardian discovered in 2018, after obtaining documents from the inspector general of the U.S. Interior Department.

 “The photographer cropped out empty space `where the crowd ended’ for a new set of pictures requested by Trump on the first morning of his presidency after he was angered by images showing his audience was smaller than Barack Obama’s in 2009,” it reported.

Had the Park Service images been sealed under a common standard, those edits also would have been available for all to see, not something that had to be uncovered through investigative digging.

President Trump’s self-proclaimed “war with the media” continued throughout the last four years. As Spicer laid out in his first briefing, the media’s “deliberately false reporting” would not be an obstacle for the administration. He warned, “the President, he will take his message directly to the American people where his focus will always be.”

Our archive of this presidential transition shows the staggering end-game for this strategy. Facts journalists presented were quickly inverted as evidence of a massive conspiracy against the President. Hatred of traditional news outlets became a sign of fealty to the President.

By day 64 in the presidential transition, events came to a head.  

Standing before the White House at a Save America Rally on January 6, Trump declared: 

“Our media is not free. It’s not fair. It suppresses thought. It suppresses speech, and it’s become the enemy of the people. It’s become the enemy of the people. It’s the biggest problem we have in this country.”

Then with Trump’s encouragement, his supporters marched on the U.S. Capitol.

The object of the rioter’s ire was not just legislators but also the press he demonized. The mob targeted journalists and destroyed their equipment. Reuters’ Editor In Charge in Washington, Jim Bourg who was at the event filing photos, tweeted: “The attitude of this crowd towards the media was very very different yesterday. I would say I heard more verbal abuse and ignorant comments about journalism and how it works yesterday than in the last 20 years combined.”

As the crowd surged into the halls of the Capitol building, “Murder the media” was carved into a doorway.

Members of the media work as pro-Trump protesters rally, January 6, 2021 REUTERS/Ahmed Gaber

The phrase "Murder the media" is written on a door to the U.S. Capitol a day after supporters of U.S. President Donald Trump stormed the U.S. Capitol Building in Washington, U.S., January 7, 2021. REUTERS/Erin Scott

Authenticity and Credibility

No one could have predicted that Spicer’s first press briefing attacking the veracity of the photos of inauguration crowds would culminate with a violent assault on the U.S. Capitol four years later, but the two events share a dangerous line of thinking.

They demonstrate how the authenticity of photography is critical for the credibility of journalism, and how the absence of it can lead to chaos.

Perhaps no better image of the consequences of this reality was the scene from the 2021 inauguration. The ceremony was already going to be smaller to allow for social distancing and protect against the spread of COVID-19. But following the violence on January 6, the crowd became even smaller yet. The entire National Mall was closed to the public, and the building was ringed with fences and razor wire.

The unprecedented security measures restricted photographer access to the Washington Monument to only a brief session during a rehearsal a day before the inauguration. A side-by-side comparison photo of the exact moment of the inauguration crowd to those in years past was therefore impossible.

But by then, no one questioned that the crowd was smaller than Trump’s inauguration, nor that the crowd was smaller by one more, as Trump became the first president in 152 years to skip his successor’s inauguration.


The side-by-side image Reuters ended up publishing this year, one of those in our prototype archive, was decidedly different.

Two photos from the same vantage point taken two weeks apart:

Original Photos

Combination picture from January 6 and January 20, 2021 show supporters of U.S. President Donald Trump fighting to enter the West front door of the U.S Capitol and President-elect Joe Biden and his wife Jill Biden walking out the West Front door of the Capitol as they arrive for his inauguration as the 46th President of the United States in Washington, U.S., January 20, 2021. REUTERS/Shannon Stapleton (Top) and Jonathan Ernst (Bottom)

Comparison slider

We built this prototype to test a new approach that is starting to take shape. As journalists did their work, each day, for 78 days —the entire transition period— their photographic records were sealed with various open-source methods. The authenticity of the work was preserved and accessible.

Our goal was to open a conversation about how new imaging technology could change journalism as we know it.

While the administrations have now changed, the issue of authentication for photos and videos remains with us. Just consider this Washington Post fact-checking story that shows how a video from the inauguration was edited to mislead people into believing that President Biden requires words be fed to him via an earpiece.

“Conspiracy theories questioning President Biden’s mental acuity and fitness aren’t new,” the article stated. “Yet as the country continues to process the events leading to and proceeding from the Jan. 6 Capitol insurrection that left one police officer and four rioters dead, attempting to deter baseless conspiracy theories becomes a higher-stakes issue.”

Looking ahead, we view this archive of the inaugural transition as a map to guide us onto a path that could restore trust in journalism. It doesn’t prescribe one route, but rather marks and affirms that we can reach this destination. By using open-source tools and methods, we can get there —together. By bringing the most diverse voices in media we can build a robust understanding of facts from multiple angles.

There is a reason for such optimism. The same surveys that point to the collapse of faith in the media show that seventy-five percent of adults believe it is possible to improve confidence in the media.

For the Starling Lab, the innovations we examined, developed, and deployed with this prototype seek to address this fundamental opportunity that newsreaders are demanding. We believe that most news consumers want the context to gauge every story they read and every photo they take in.

The alternative to that is unthinkable.