If you sign up for Twitter’s premium service, you unlock access to Grok, the AI bot on the platform, and a trending news feed in the Explore tab powered by artificial intelligence. However, there appears to be a notable issue concerning the quality of this particular feature.
The operation of Grok’s trending news feed is relatively simple: it gathers the most popular posts related to a specific news event and then generates a news summary based on those posts. While the use of generative AI for this purpose is not groundbreaking, it is crucial to assess the accuracy of its reporting process before contemplating replacing human writers and editors with Grok.
It seems that Grok is aggregating humorous tweets and presenting an AI-generated output as authentic news. This was highlighted in a post by user BrettRedacted following the recent earthquake in the New York City metro area. The bot crafted a headline that read, “Adams vs. Earthquake: 50,000 Cops in Subway Showdown,” implying that Mayor Eric Adams is mobilizing the NYPD to “prevent further earthquakes,” considering the deployment of “robo cops,” and directing “every cop in the city” to “shoot the damn earthquake before it strikes again.”
While we recognize the unconventional nature of the present times, it is highly improbable for any sensible individual to view Grok’s news summary as accurate. In case of any skepticism, one could examine the original posts contributing to that news summary, which, in this case, were all jokes regarding the mayor’s response to the earthquake.
This situation epitomizes the current landscape on Twitter in 2024. Previously, the platform was celebrated for delivering real-time updates on significant events such as the NYC earthquake while also offering entertainment with humorous perspectives on the situation. However, the current scenario signifies a shift where jokes are misinterpreted as news, emphasizing the saying “you get what you pay for.”
It is not difficult to envision this scenario evolving from being sadly amusing to potentially hazardous. What if Grok chooses to “report” on a seemingly legitimate topic that is actually based on the rampant spread of misinformation across the platform? Users could be deceived into believing false narratives, like their city being under siege or a candidate being involved in illicit activities, due to the misinformation presented as news.
While we may not have the ability to prevent the platform from endorsing such content, we can collectively decide not to view Grok or any generative AI as a reliable news source or an accurate summarizer of daily events.
