news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

July 18 news, ElonMuskIn the future vision envisioned by Elon Musk, people can get news information on the X platform through its artificial intelligence model Grok. However, this vision has recently faced practical challenges, exposing the limitations of current artificial intelligence technology.

Former U.S. President DonaldTrumpGrok quickly generated multiple misleading headlines during the sensitive period of the assassination attempt on Donald Trump. One of the false reports claimed that Vice President Kamala Harris wasShooting, the mistake may have originated from a sarcastic comment by a user on the X platform about a previous incident in which President Biden confused Trump's name with Harris.

Another Grok-generated news feed incorrectly identified the shooter and falsely linked him to the far-left network Antifa. Officials later released the suspect's real name and have yet to determine his motive.

Grok's content essentially transforms online jokes, rumors and misunderstandings into news reports on the platform. Although each news summary is accompanied by a disclaimer: "This article is a summary of posts on the X platform, and the content may change over time. Grok may make mistakes, please verify its output yourself."

Musk has been promoting his social media platform X and using Grok's AI capabilities to automatically generate news headlines and summaries based on hundreds of millions of users' posts. He has repeatedly criticized traditional news media for being inefficient and unreliable, and encouraged X users to check Grok regularly for news.

"What we're doing on X is aggregation," Musk said at an advertising industry conference. "We're using AI to combine the intelligence of millions of users, and I firmly believe that this will become a new model for news dissemination."

X Platform has not yet responded to questions about the accuracy of Grok's title.

Grok is a product of Musk's artificial intelligence company xAI. It began to develop and gradually launch new features last year, including providing chatbot services for X platform users.

Grok has demonstrated its accuracy in some news summaries, showing the potential of the technology. Musk said that Grok's ability to access posts on X in real time is a major advantage, and it is seen as an "AI search assistant with a slight sense of humor."

However, recent mistakes have also shown the potential weaknesses of such humor-oriented models when they try to process a huge volume of posts from a variety of sources in real time.

Katie Harbath, Facebook's former head of public policy, noted that despite the progress, humans are still needed to provide necessary context and fact-checking when dealing with breaking news such as shootings.

Even experienced journalists can make mistakes when handling breaking news, but Grok's blunder went beyond the usual confusion.

Another example of Grok-induced misunderstanding is surfacing, with the title: "Home Alone 2 actor shot at Trump rally?" This claim actually originated from Trump's brief cameo in the movie "Home Alone 2" and the discussion of the movie among users on X. Unfortunately, when Grok generated this title, it did not clarify whether the so-called "actor" referred to Trump himself.

Generative AI tools often struggle with accuracy in their quest for efficiency and innovation. For example, Google recently corrected its AI capabilities, which produced illogical results in some cases, such as suggesting that the cheese on a pizza should be glued in place.

Faced with these challenges, some companies have chosen a different route from Musk and have not positioned their products as real-time news tools. For example, OpenAI's ChatGPT responded to queries about Trump's attack by providing a summary of the event and linking to fact-checking sites such as PolitiFact, while explicitly stating that it is "not a real-time news product."

Meanwhile, Meta executives said they do not plan to promote political content on its Threads platform to avoid the negative impact of politics and hard news.

Over the weekend, users complained that Threads was slower than X in reporting on the shooting, showing X’s advantage in news dissemination.

In an internal email obtained by The Wall Street Journal, X CEO Linda Yaccarino sent a directive to all employees on Sunday calling the attempted assassination a "significant incident" and pledging to prioritize posts that violate the platform's rules, such as content that praises the attack.

It is worth noting that this is not the first time that Grok has been challenged in summarizing news events. After the presidential debate in June this year, Grok generated the wrong headline "Newsom Wins Debate" due to misunderstanding the discussion about California Governor Gavin Newsom. In fact, Newsom did not participate in the debate, but only attended the event.

Previously, Twitter had a dedicated team responsible for manually compiling summaries of trending topics. But former Twitter executive EvanHansen(Evan Hansen) revealed that the team was disbanded during the massive layoffs after Musk took over.

Hansen also said that Twitter had begun trying to use AI to write summaries before Musk took over, but human review of the content was still required. He pointed out: "AI can play a role, but it must be handled very carefully to ensure the quality and accuracy of the content." (Xiaoxiao)