Elon Musk’s X Chatbot Grok Spreads False Iran-Israel Attack Headline

Artificial Intelligence, News, Social Media

On Thursday, Elon Musk’s X platform sparked controversy by prominently featuring a fabricated news story on its main feed. The headline claimed that Iran had launched an attack on Israel, which would have been a deeply troubling escalation in an already tense region.

You can open Table of Contents show

The story seemed plausible, given that Israel had conducted an airstrike on Iran’s embassy in Syria earlier that week, resulting in the deaths of two generals and other officers. Retaliation from Iran appeared to be a real possibility.

However, the shocking truth soon came to light: the headline was entirely fake. Iran had not attacked Israel. Worse, on the very first day of an updated version of the feature, X’s own official AI chatbot, Grok, generated the false headline, which the platform’s trending news product, Explore, then amplified. This incident raised serious concerns about the spread of misinformation and the role of AI in news generation and dissemination.

Press the search button 🔍 to see real-time customized news for you created by Grok AI

— Elon Musk (@elonmusk) April 5, 2024

How did Musk Allow this to Happen?

Before Elon Musk took over, Twitter introduced a new feature a few years ago to give users more context about trending topics. Instead of relying solely on algorithms, Twitter decided to bring in the human touch. They assembled a team of editors tasked with curating news and offering written descriptions to help users understand why certain topics were gaining traction.

Before this change, Twitter’s algorithms would simply identify popular keywords and phrases, making trends visible to users. However, the reason behind a trend’s popularity wasn’t always clear. That’s where the human editors came in. Starting in 2020, they worked to put trends into perspective, giving users a better grasp of the stories unfolding on the platform.

But Twitter didn’t stop there. In 2021, they partnered with two trusted news agencies, AP and Reuters, to further enhance the credibility and quality of the context provided for trending topics. Due to this collaboration, the top tweets displayed under a trend were no longer just algorithmically ranked but also carefully selected by human editors.

However, when Elon Musk acquired Twitter in October 2022, things took a turn. Shortly after the takeover, the written context that had become a valuable feature for users suddenly vanished. As it turned out, Musk had let go of Twitter’s human editors as part of a series of layoffs. In November 2022, Reuters reported the axing of Twitter’s curation team, responsible for ‘highlighting and contextualizing the best events and stories that unfold on Twitter’.

The disappearance of human curation on the platform raised concerns about the potential spread of misinformation and the lack of context surrounding trending topics. It served as a reminder of the important role human judgment plays in the age of algorithms and the rapid dissemination of information on social media.

X’s Updated Explore Page

A few X fan accounts, such as @XDaily, released screenshots of the impending revamp of X’s Explore page earlier this week. The platform’s notorious trending topics list is included on X’s Explore page, along with area breakouts like “Sports” and “News” that show users the top stories in each niche.

DEMO: Updated demo of the upcoming explore page.

pic.twitter.com/4KeFuIyZpT pic.twitter.com/tBGiGxBP9n

— X Daily News (@xDaily) April 3, 2024

When Elon Musk revamped the Explore page on X (formerly Twitter), it seemed like a step towards bringing back the much-needed context for trending topics and stories that had been missing since he took over.

The updated Explore page, which started rolling out on April 4, offers a more comprehensive look at what’s trending. Now, when you click on a trending topic, you’ll see top posts from users, giving you a sense of the conversation happening around that subject. But that’s not all—above the user content, there’s a concise summary that provides an overview of the topic at hand. It’s like having a TL/DR right there for you.

But the real game-changer is at the very top of the page. X has started creating headlines for each trending topic, presenting the information in a way that feels like you’re about to dive into an article. It’s a smart move that makes the trending stories feel more substantial and easier to engage with.

And here’s the kicker: the Explore page has gotten a major upgrade in terms of visibility. Previously hidden, the trending stories now occupy a prominent position on the right-hand sidebar of the main X homepage. You know where the trending topics list used to be? Yeah, that’s where you’ll find Explore now. And with hundreds of millions of people checking out X every day, that’s a lot of eyeballs on those trending stories.

All in all, it seems like Musk and his team at X are making a solid effort to bring back the context and curation that made the platform’s trending topics more meaningful and easier to understand. It’s a step in the right direction, and I’m curious to see how it evolves from here.

But Musk hasn’t employed additional human editors to provide the context that X currently provides, nor has he brought back Twitter’s curation crew.

X’s AI chatbot Grok writes the context.

X’s AI Dependence Risk

It’s no secret that Grok, the AI feature that Elon Musk has been pushing hard on X, is the driving force behind the platform’s contextualized summaries.

Groooook https://t.co/kuJE4XhJdq

— Elon Musk (@elonmusk) November 5, 2023

X makes it clear as day, with a little note at the bottom of every Explore page that reads, “Grok is an early feature and can make mistakes.””Verify the outputs.”

Well, on Thursday, Mashable got a front-row seat to one of those mistakes. We watched as the “Iran Strikes Tel Aviv with Heavy Missiles” story started trending on X, and let me tell you, it was a doozy.

Turns out, the topic gained traction because a bunch of blue checkmark accounts (you know, the ones that pay X a monthly fee for premium features like that shiny verification badge) started spamming the same copy-paste misinformation about Iran attacking Israel. The posts that X curated for the story were chock-full of verified accounts spreading fake news, along with an unverified video showing some explosions.

From what we can tell, X’s algorithms picked up on the potential story trend within these users’ posts and created an Explore story page. And since X has been pretty open about how things work behind the scenes, it’s safe to assume that Grok whipped up an official-looking written narrative and a snappy headline to go with it.

The catch? It did all this based on a select group of users sharing bogus information, in an attempt to automatically provide context for what the platform itself thought was a legit story.

Now, this isn’t the first time Grok has served up a hefty dose of misinformation. Previous reports on early versions of X’s chatbot found that it often cooked up fake news in private chats with the lucky few who had access to it.

However, this recent incident, when paired with the new Explore feature, represents the first instance where X disseminated Grok’s misinformation to its entire user base, presenting it as a legitimate trending news story and offering context for an actual event.

Since Musk took over, disinformation on X has been spreading like wildfire. And thanks to the new Explore product, the company ended up promoting a straight-up lie spread by paying bad actors who stand to make a buck by generating engagement.

Just one day after Grok generated this fake story that X then promoted through its Explore page, the company rolled out Grok to all Premium-subscribed X users, giving them access to the misinformation-creating AI chatbot as well.

The experience is exhilarating, demonstrating that despite its advanced features, AI can still make grave errors.