Why The Washington Post is inviting sources to annotate stories

Last week The Washington Post announced a new system, called “From the Source,” that will let sources annotate stories they appear in after the stories are published. Readers will see the annotations if they click or hover over a source’s name in the story, and the annotations will also appear in a separate section of the story’s comments.

It’s the latest in a series of changes the Post has been making in an attempt to increase engagement on its site. About a year ago it introduced a chatbot for climate questions and later expanded it into a tool called “Ask the Post AI,” which draws from Post stories across the newsroom. It injected AI into the comments in December, generating prompts and summaries of the discussion¹, and the Post homepage now also includes a ticker of headlines from other publications, allowing readers to skim those headlines without leaving the Post’s site.

It remains to be seen if any of these changes will attract new subscribers or bring back any of the hundreds of thousands of people who canceled their subscriptions because they disapproved of changes to the Post’s opinion pages mandated by billionaire owner Jeff Bezos; those subscribers presumably canceled because of a difference in ideology rather than a lack of avenues for engagement.

From the Source will first be tested on the Post’s climate coverage, much like the chatbot was. “The climate team has been a lab for audience experiments over the past few years,” said Zachary Goldfarb, editor of the Post’s Futures section, which combines the newspaper’s tech, business, health, climate, and science teams. The Post dramatically expanded its climate team in 2022, and has since been using the climate vertical as a testbed for ideas that eventually make their way to the rest of the newsroom.

“The Post is going through a period of reinvention and trying out a whole bunch of experiments to engage readers and capture attention, and this is sort of the latest experiment on that front,” Goldfarb told me.

Here’s how the annotations will work: After a story publishes, reporters will send each named source a special link that lets them read the story and add a comment. After a source submits annotations, a member of the audience team will review and decide whether to publish them. They’ll also loop the reporter back in if they think the annotation is worth publishing but deserves a response; in such instances, the source’s annotation and the reporter’s reply will appear next to each other.

The first article to include source annotations, a story about reusable cups at concert venues, ran on Monday. When I first read the story a day after it published, it included a four-paragraph annotation, largely about the benefits of reusables, from one source who runs a reusable packaging company. A few hours later, an additional annotation appeared from another source, who runs yet another reusable packaging company. That one, just a paragraph long, mostly plugged that source’s company.

When the Post announced From the Source, the reaction online was swift and scathing. Many people, including members of Nieman Lab staff, wondered whether the annotations might give a platform to sources who might want to undermine the reporting, spread misinformation, or simply promote their own political or economic interests. Goldfarb said he’s aware of the criticism — the reaction within the Post’s newsroom itself, he said, has been “wide-ranging,” and journalists are worried about giving up control of a story — but he also hopes the initiative will increase transparency and build trust by breaking down the walls between sources, readers, and reporters.

“It’s very much a pilot project right now,” he said. “We’re seeing how it goes, and we’ll learn from that and make adjustments along the way. I understand people’s concerns, and they’re legitimate, but people can already share misinformation in our comments and on social media. We’re basically bringing some of that to our platform and engaging directly with it. We’re not ignoring the reality of that conversation that’s already happening, and we want to do so in a responsible manner. I think there will be some close calls, and we often make very close calls in journalism. But we’re not giving up our fundamental framework where we make assessments about what we should publish.”

The best-case scenario, Goldfarb said, is one where sources are able to provide useful context that wouldn’t otherwise make it into a story. “We know that climate scientists take public communication very seriously and often want to engage and provide additional layers of feedback,” Goldfarb said. (This is something I can confirm from my own experience as a climate journalist.) “If you look at social media, you see great discussions related to climate change on multiple platforms. So I think it makes it a good place for offering that kind of additional discussion.”

1. According to the AI summary of comments on the story announcing the AI summaries, “The comments overwhelmingly express dissatisfaction with recent changes at The Washington Post, particularly the introduction of AI features. Many readers feel the changes are unnecessary, complicate the user experience, and suppress genuine dialogue. There is a strong sentiment that these updates alienate subscribers and detract from the newspaper’s value, with some commenters indicating they plan to cancel their subscriptions.”