When algorithms come for journalists

Journalists — especially those without institutional newsroom support — rely on tools from major tech companies like Google and YouTube for newsgathering, production and distribution as a matter of course. As these information giants publicly wrestle with controversial content moderation decisions that dominate headlines and Congressional hearings, their decisions also run the risk of stifling routine reporting. When content is removed or an algorithm tweaked behind closed doors, news organizations and journalists are often left without any sort of transparency into the process or a clear path to appeals.

In the last month, Freedom of the Press Foundation and the U.S. Press Freedom Tracker experienced this first-hand, with the temporary takedown from Google Docs of an online database we've used to track more than 2,500 tweets by former President Donald Trump attacking the media. We've used this public spreadsheet for data analysis over the years, and provided it to readers and other journalists to do their own exploration of Trump’s anti-media tweets and their effect on press freedom.

Two weeks ago, the database was taken down by Google and replaced with a notice claiming an unspecified “terms of service” violation.

Neither Freedom of the Press Foundation nor Stephanie Sugars, our reporter whose personal account "owns" the document, were notified of its removal, and no recourse was offered.

“When I discovered the spreadsheet had been flagged, I was at a complete loss for how to contest the decision. Even the directive from Google's Help screen was useless, as it said to request a review when the document was already open. I couldn't open the document at all, and when I tried it disappeared from my Drive,” according to Sugars.

Still, we were lucky. Some of our colleagues know employees at Google, and our allies know even more. After many people made private inquiries on our behalf, the document was restored without explanation a day after we discovered it was down. Obviously, that course of action is not available to most.

We still have no idea why the Trump tweet database was taken down.

Even a temporary suspension can have serious drawbacks for reporters or outlets that are providing timely reporting or live broadcasts. Since the beginning of 2021, for example, the progressive news outlet Status Coup has seen both widespread adoption of its live-streaming footage and an increase in restrictions from YouTube, where it broadcasts. As it was covering a pro-gun rally in Virginia last month, its feed was abruptly cut for violating the service's firearms policy. Similar to allies stepping in in our Trump tweet database situation, high-profile criticism of Silicon Valley appeared to lead to the stream’s restoration.

That option is not available to most. Should local journalists have to rely on higher-profile journalists to draw attention to their case?

And despite that reversal, Status Coup continues to face problems. Even though their channel was restored, Status Coup has made the editorial decision not to "go live" as it covers certain controversial events. Critical footage it shot from the Capitol Riots, which was later licensed by CNN and other networks and seen by millions of people, was taken down by YouTube, and much of it has not been restored. In the weeks since, some of the same raw footage was also removed from Google Drive, again citing unspecified Terms of Service violations, according to the outlet’s co-founder Jordan Chariton.

Chariton also described the difficulty of planning reporting trips or assignments when the resulting stories are shaded with total uncertainty. “As a journalist you want to sink or swim on your judgment, what stories you choose, the way you report, building relationships with sources,” he said, but Status Coup’s experience thus far “shows that you can be tenacious, work seven days a week, break big stories, and Google and the rest of them could choose to bury you, choose to take your footage down.”

Google’s domination of search means it can have profound effects on distribution even for outlets that don’t expressly rely on products like YouTube. U.S. Right to Know, a non-profit newsroom that engages in investigative journalism on public health issues, has provided Freedom of the Press Foundation with evidence of a sudden and dramatic drop-off in incoming traffic from Google search results after the search engine released a “core update” to its ranking algorithm.

U.S. Right to Know’s Google referral traffic dropped off a cliff in the beginning of December, right when the algorithm change was announced. Was there some action that U.S. Right to Know took that triggered this? What can it do to rectify the situation? Its editors have no idea.

U.S. Right to Know has also previously attracted strong negative attention from the subjects of its reporting — Monsanto set up an "intelligence center" to monitor and discredit the organization and other journalists. In the absence of meaningful transparency or an appeal process, it's difficult to rule out the idea that such a motivated company could possibly have played a role.

This is only one example of many over the years where an algorithm change has killed a news outlet’s traffic overnight. The New York Times wrote in 2017 how legitimate left-leaning independent news outlets were getting decimated by Google’s attempt to eliminate “fake news.” Outlets large and small have long complained Facebook constantly wreaks havoc on their traffic — and in turn, their revenue — if they decide to de-emphasize certain subjects or news in users’ newsfeeds. Newsrooms are stuck trying to read the tea leaves in vague announcements for how to respond.

To be clear, companies like Google, Facebook and Twitter have their own First Amendment right to make decisions about what content they host and how they present that content. Efforts to diminish the liability protection afforded to tech platforms and websites of all stripes are misguided at best — and would unequivocally hurt the cause of free speech.

But that doesn’t mean the companies that operate dominant tech platforms shaping our social information intake shouldn’t be held accountable when their opaque decisions harm independent reporting and journalism.

Tech companies are so dominant that it’s impossible to make or distribute news without them for many journalists. Given the importance of a strong and independent press, it’s time these companies prioritize and allocate the resources that protect journalists from unexplained, random or otherwise punitive abuses of their power.