On April 5, 2024, fake news announcing an attack by Iran against Israel was broadcast by X.com’s AI chatbot Grok.
“ Iran hits Tel Aviv with heavy missiles », This is the information that we could read in trending news on X.com (formerly Twitter). In fact, everything suggested that this was proven information since earlier, Israel had carried out an airstrike against the Iranian embassy in Syria.
But if the information was enough to chill the blood, it turned out to be much more worrying after we discovered that it was fake news, created from start to finish by Grok, the AI chatbot from X.com.
Grok accessible to paid accounts, which relay all its information, including fake news
It all started when Grok, X.com’s AI chatbot, posted news that Israel had been attacked by Iran. This information quickly caught the attention of users, who began sharing the information on various social media platforms. However, there was no evidence of such an attack in mainstream news or government sources. So how did this fake news become so viral that many users believed it?
Quite simply because of how Grok works. As a reminder, this is a tool developed internally by X.com with the aim of powering its new “Explore” functionality dedicated to trends. Thus, without any human verification, it is indeed artificial intelligence which completely invented and wrote this misleading title.
You should know that in the past, Grok had already generated false information in private conversations with certain users. But this time, it was the Explore functionality that promoted this misleading content on the platform’s home page. And even if, by his “own” admission, Grok “ can make mistakes “, the damage was done, paying users, who were able to have exclusive access to Grok, jumped headlong into the trap of fake news and relayed it. And since it is no longer possible to report fake news on X.com, anything seems permitted.
How did we get here?
To understand how this happened, we need to go back a few years. Before Elon Musk acquired Twitter, the platform implemented a feature that provided written context for trending topics. This feature relied on human editors who organized news and provided context to trending topics. However, shortly after Elon Musk purchased Twitter in 2022, he laid off human editors, leading to the disappearance of written context on trending topics.
X.com’s “Explore” page, which was recently updated, appears to bring back the written context of trending topics. The page indeed includes a list of trending articles in different subject categories. It provides the main user articles on each topic accompanied by a short summary. At the top of the page, a headline created by X.com presents the information to readers as if they were about to read an article. But… no human is behind this information sharing interface.
And for good reason, Elon Musk did not bring back Twitter’s curation team or hire new human editors. Instead, X.com’s AI chatbot, Grok, is now responsible for writing the context provided on the “Explore” page. The highlighting and relaying of this fake news are proof of the dangers of using AI at all levels of the Internet. And to think that Elon Musk previously wanted to demonetize fake news on X.com…
Source : Mashable
7