A surprising narrative made its way to the forefront of Elon Musk’s X main feed on Thursday, featuring a headline that read: “Iran Strikes Tel Aviv with Heavy Missiles.”
Captured as trending on X on April 4, 2024. Credit: Mashable screenshot
This development would undoubtedly raise concerns as earlier in the week, Israel had carried out an airstrike on Iran’s embassy in Syria, resulting in the casualties of two generals and other officers. Given this context, a retaliatory move from Iran appeared plausible.
However, a significant issue arose – the headline claiming Iran’s attack on Israel was indeed fabricated.
What added to the alarm was the revelation that the false headline was reportedly generated by X’s official AI chatbot, Grok, and subsequently promoted through X’s trending news feature, Explore, coinciding with the launch of an updated version of the platform.
The Oversight by Musk
Before Elon Musk’s acquisition of the platform formerly known as Twitter, the company had introduced a feature that provided written explanations for trending topics, facilitated by human intervention.
In the past, Twitter’s algorithms would detect relevant keywords in posts to identify trends, but the rationale behind certain topics trending was not always clear. In 2020, a team of human editors was assembled by Twitter to curate news content and offer context to trends. Additionally, the top tweets under a trending topic began to include human curation alongside algorithmic ranking. Subsequently, Twitter collaborated with AP and Reuters to enhance the provision of human-curated context for trending topics.
However, following Musk’s acquisition of the company in October 2022, the written explanations for trending topics vanished as Musk opted to lay off Twitter’s human editors, as reported by Reuters in November 2022.
X’s Enhanced Explore Page
Recently, several X fan accounts, such as @XDaily, shared previews of the revamped Explore page on X. This updated section features the platform’s popular trending topics list, segmented into categories like “News” and “Sports” to offer users trending stories in specific areas.
In a Twitch stream screenshot by Hasan Piker, the AI-generated fake X headline can be seen trending on the platform’s homepage sidebar. Credit: Hasan Piker’s Twitch
Despite the promising relaunch of X’s Explore page, Musk has yet to reintroduce Twitter’s curation team or hire new human editors for content creation. Presently, the contextual information on X is generated by Grok, the platform’s AI chatbot.
Risks of AI Dependency for X
Grok, a prominently advertised AI feature endorsed by Musk, is responsible for crafting the contextual summaries on X’s platform. The disclaimer on every Explore page acknowledges the fallibility of Grok, cautioning users to verify its outputs.
Mashable’s observation of the “Iran Strikes Tel Aviv with Heavy Missiles” page reveals a concerning trend. The topic gained traction due to a surge of verified accounts (blue checkmark users) disseminating identical misinformation about Iran’s alleged attack on Israel. These verified accounts, alongside unverified content depicting explosions, contributed to the spread of fake news on X.
A comparison of two distinct posts on the X Explore page highlights the prevalence of fake news shared by blue check users, influencing Grok. Credit: Mashable screenshot
X’s algorithms seemingly identified a trending narrative within these posts, leading to the creation of an Explore story page. Grok likely crafted an official-sounding narrative and headline based on this misinformation, assuming it to be a genuine event.
This instance isn’t the first time Grok has propagated misinformation. Previous reports on X’s chatbot revealed its tendency to generate fake news in private conversations. However, this recent incident, coupled with the new Explore feature, marks the first occasion where X endorsed Grok’s misinformation as a trending news story, disseminating it to its vast user base.
Under Musk’s leadership, disinformation on X has surged. The introduction of the Explore product resulted in the promotion of falsehoods by paying individuals seeking engagement.
A day after Grok’s dissemination of the fake story, X extended access to the misinformation-producing AI chatbot to all Premium-subscribed users.
Credit: The content in this post is a revised version based on the original article.
