Elon Musk sent me an email this week with some surprising details about his plan to distill and present news on X using AI. I had contacted him after trying Grok — X’s AI chatbot — and noticing it didn’t link to a Time story it summarized. I wanted to click into the article and read more, so I reached out.
Musk mentioned that better citations are on the way but shared a more profound vision for the product, which he aims to develop into a real-time compiler of news and social media reaction. Essentially, his plan is to utilize AI to merge breaking news and social commentary around major stories, present the compilation live, and enable you to delve deeper via chat.
“As more information becomes available, the news summary will update to include that information,” Musk informed me. “The goal is simple: to provide maximally accurate and timely information, citing the most significant sources.”
Accomplishing that goal won’t be easy, but the bot might become an innovative news product given its access to the X firehose. “Grok is analyzing sometimes tens of thousands of X posts to render a news summary,” Musk stated.
Currently, Grok is displaying a running list of headlines and integrating social reaction into its summaries, including the discussion around the Time story I was interested in regarding Trump’s potential second term .. Grok has plenty There is room for improvement, and they will need to address issues such as citation and accuracy, but it could be valuable if X gets the execution right.
“That’s actually what I used to come to Twitter for — news and commentary,” Ben Smith, editor-in-chief of Semafor and author of Traffic, told me.
Discussion on X will form the foundation of Grok’s summaries — or, essentially all of it. Musk mentioned that Grok will not look directly at article text and will instead rely solely on social posts. “It’s summarizing what people say on X,” he said. “It definitely needs to do a better job of displaying relevant posts, including, for example, the TIME post that links to the article.”
Musk’s approach will differentiate Grok from all AI news summarizers, and likely be more contentious, but there is an opportunity to satisfy users, publishers, and the platform together. It starts with solid citation, providing users with a way to delve deeper into the source material once their curiosity is piqued.
Josh Miller, CEO of The Browser Company, whose Arc browser is doing AI summarization, told me that platforms, users, and publishers could all win with aggressive citation. “People want trust in the data,” he said. “They want to understand where it’s coming from. And more importantly, if we do a good job giving them what they want, they want more of it. They get more curious. They want to read deeper.”
Grok could similarly be an entry point to news stories that people might otherwise not see, giving them more reason to come back and engage. “I definitely don’t go to Time.com,” Miller said. “So if that was pushed to me, I would probably go read that Time article. And I wouldn’t have seen it otherwise.”
Igor Babuschkin, a technical staff member working at Musk’s xAI, said his team is focused on “making Grok understand the news purely from what is posted on X.” Regarding citation, he said that “since news is often discussed on X, this can sometimes lead to Grok making references to existing news outlets. We are working on improving the citations so that we reliably capture who the information in the article comes from.”
For publishers, the value exchange on social media has always been problematic, and this time is no different. News companies' work, like the Time article, influences news cycles. Summarizing that work without a prominent link back could disconnect them from the economic reward, hurting their ability to create more. However, publishers are shifting away from relying solely on social media traffic, which has always been uncertain, even though the consequences have been severe for the industry.
"As an editor in this essentially post-Twitter news environment, I would prefer the platform took my content instead of taking my journalists' time," said Semafor's Smith.
Creating summaries only from the posts on X, and not the articles themselves, could create some distance between X and the publisher lawsuits targeting AI companies. Eight newspapers sued OpenAI and Microsoft for copyright infringement this week, for example. sued OpenAI and Microsoft for copyright infringement this week, for instance. And The New York Times has sued OpenAI and Microsoft as well.
Courts usually consider commentary under the provision for fair use, as opposed to original work, said Danielle Coffey, president and CEO of the News/Media Alliance, which represents thousands of news publishers. However, using only commentary to summarize the original source, she said, "is a bit too cute by half though.
Under Musk, X seemed to initially lower the priority of news by temporarily removing headlines from articles, revoking verifications, and making some initial algorithm changes, but Musk's emails show that he still sees news as a fundamental X capability. It's a potential competitive advantage as competitors like Meta's Threads distance themselves from news.
In an ideal world, original sources and commentary could blend in Grok, drawing details from insiders, analysis from outsiders, and commentary from the herd. Leaders at Twitter had long envisioned such a scenario. And perhaps the latest advancements in generative AI have made it worth attempting once more.
This article is from Big Technology, a newsletter by Alex Kantrowitz.
