On Thursday, a shocking story was promoted on Elon Musk X’s “front page,” or main feed:
The headline read “Iran attacks Tel Aviv with heavy missiles.”
Captured the trend on X on April 4, 2024.
Image source: Mashable screenshot
This is certainly a concerning world news development.Earlier that week, Israel implement An airstrike on the Iranian embassy in Syria killed two generals and other officials. Iranian retaliation seems likely.
I gave Elon Musk’s X a mandatory blue check. This is what it looks like.
However, there is one major problem: Iran did not attack Israel. The title is fake.
Even more worryingly, the fake headline was apparently generated by X’s own official AI chatbot, Grok, and then promoted by X’s popular news product, Explore, on the first day of an updated version of the feature.
Tweet may have been deleted
How Musk made this happen
A few years before Elon Musk acquired the platform then known as Twitter, the company launched a new feature that offered written context For hot topics. Technology to power written trend descriptions? Humanity.
Previously, when a story happened, Twitter’s algorithm would capture relevant keywords posted and share the trend, but sometimes it wasn’t obvious why a given set of words became trending.
However, in 2020, Twitter assembled a team Human Editor Curate news and understand trends. In addition, popular tweets that appear under trending topics will no longer be ranked only by algorithms, but also by manual curation.A year later, Twitter partnership Partnering with The Associated Press and Reuters to further strengthen its efforts to provide human context for trending topics.
However, written background on the hot topic disappeared shortly after Musk acquired the company in October 2022. Musk fires Twitter’s human editors. As Reuters reported in November 2022, “Twitter’s curation team, which ‘highlights and contextualizes the best events and stories happening on Twitter,’ has been laid off.”
XNewly updated exploration page
Earlier this week, some X fan accounts such as @XDaily shared screenshot X Upcoming Discovery Page Restart. X’s Explore page includes the platform’s infamous list of trending topics as well as specific sections like “News” and “Sports,” giving users the top stories in each specific topic category.
Tweet may have been deleted
Musk’s updated Explore page looks poised to bring back written context for trending topics and stories.
On April 4, X began rolling out a newly updated Explore page. The update provides top user posts for each popular topic as well as easy-to-understand summaries of topics above user content. At the very top of the page, X creates a headline that presents information to readers as if they were to read an article about the topic.
Additionally, the update gives Explore a new prominent place on the platform. Explore’s top stories are now embedded directly into the right sidebar of the main X page, where the list of trending topics once stood, viewed by hundreds of millions of daily X users every day.
In a screenshot captured during popular Twitch user Hasan Piker’s live stream, an AI-generated fake X title can be seen trending on the platform’s homepage sidebar.
Image source: Hasan Piker’s Twitch
However, Musk didn’t bring back Twitter’s curatorial team or hire new human editors to write the context that X now provides.
Context is written by Grok, X’s AI chatbot.
The danger of X depends on the AI
We know that Grok is an artificial intelligence feature promotion go through Muskproviding support for X’s contextual summary because X says so in the fine print of each exploration page.
Tweet may have been deleted
“Grok is an early feature and may make mistakes,” X writes in a small note below the written context provided on the Explore page. “Verify its output.”
Mashable observed X’s “Iran attacks Tel Aviv with heavy missiles” page spreading fake news in a trend on Thursday.
From our observations, this topic started trending because blue checkmark accounts (users who subscribe for The curated posts provided by X are filled with these verified accounts spreading fake news as well as unverified videos depicting explosions.
A side-by-side comparison of two separate posts on the X Explore page shows fake news posts from blue check users that notified Grok.
Image source: Mashable screenshot
From there, X’s algorithm seemed to notice potential story trends in these user posts and created an “Explore Stories” page. We can infer from X’s own statements about its inner workings that Grok must have created an official-looking written narrative, as well as a catchy title. All it does is based on specific users sharing fake news, automatically trying to provide context for what the platform itself seems to believe is a true story.
This isn’t the first time Grok has provided error messages to users. Previous reporting on an early version of the X chatbot found that it frequently created fake news in private chats with a handful of users who had access to it. However, this recent incident, combined with the new Explore feature, marks the first time that X has packaged Grok’s misinformation into an actual trending news story and promoted it to its entire user base, ostensibly as a real event background.
Under Musk, disinformation has skyrocketing On a platform named X. With the new Explore product, the company promotes a disinformation campaign that spreads falsehoods by paying bad actors to make money by generating engagement.
one day After Grok generated the false story and X promoted it through its Explore page, the company rolled out Grok to all X users with a Premium subscription so they could also use the AI chatbot that created the misinformation.
Tweet may have been deleted