Elon Musk has ‘fixed’ Grok—to be more like him than ever

As Elon Musk announced plans over the Fourth of July weekend to establish a third political party, his AI chatbot, Grok, gave a sneak preview of the information ecosystem in that party.

It’s pretty much indistinguishable from Musk’s personal feed on X.

Despite pledging just two months ago to dramatically reduce political spending in the future, Musk appeared so incensed by the passage of President Trump’s tax bill last week, and its potential impact on the U.S. deficit, that he vowed to start his own party. The announcement sent shockwaves through the media landscape, even during a tragically eventful holiday weekend, prompting a livid response from the president and a sharp drop in Tesla shares on the market.

It remains unclear for now whether Musk has put much thought into the complex mechanics of actually launching a new political party, or if he instead thinks that establishing one is as simple as Steve Carell’s character from The Office attempting to declare bankruptcy ">just by shouting. Something the Tesla tycoon seems to have considered more carefully, though, is how Grok might function in an election featuring a Musk-approved third-party candidate.

Going into the holiday weekend, he announced that a new and improved Grok had landed on X.

The difference in the AI chatbot was perhaps most pronounced in its response to a question from an X user about whether “electing more democrats” would be “a bad thing.” Grok replied in the affirmative, claiming that “electing more Democrats would be detrimental, as their policies often expand government dependency, raise taxes, and promote divisive ideologies.” 

In order to back up this broad, generalizing claim, Grok cited infamous right-wing think tank the Heritage Foundation, rather than any news source. (That particular Grok reply has since been deleted, though several X users have now asked similar questions and received similar replies, with the definitive word would softened to the more speculative would likely or could be.)

Obviously, in an election year this kind of messaging would fire up, ahem, would likely fire up Musk’s fan base, which includes more than 221 million followers on X. However, it would likely also make good on the tech titan’s previous pledges to “fix” Grok.

A bot is born

Grok was designed with the explicit intention of being an “anti-woke” competitor to OpenAI’s ChatGTP, which Musk cited as being compromised. (The feud between Musk and OpenAI CEO Sam Altman stems back to the days when Musk sat on OpenAI’s board, which he stepped down from in 2018.) Musk seemed to want Grok to become the arbiter of truth for X users, so much so that “Grok is this true” has since become a frequent meme on Bluesky to comment on news that might trouble ardent MAGA supporters. 

From the moment Musk debuted Grok in late 2023, though, conservative X users have complained that, despite Musk’s best efforts in training it otherwise, the chatbot still appeared to suffer from the woke (artificial) mind virus. (“Grok will get better,” he promised at the time.)

In the year and a half since, the chatbot has supplied some startling examples of what Musk’s idea of “better” might entail. In July 2024, when then-President Joe Biden announced he would no longer seek reelection, for instance, Grok offered electoral disinformation about the legality of switching candidates at that point in the race, prompting five secretaries of state to send Musk a letter demanding he “immediately implement changes to X’s AI search assistant, Grok, to ensure voters have accurate information in this critical election year.” 

Now, in the year after that critical election, Grok has evinced some of Musk’s notorious animosity toward “legacy media,” along with his opinions about the phrase white genocide. The flashpoint that inspired Grok’s most recent update, however, seemed to have arrived last month, in the wake of a political assassination in Minnesota.

Working on it

On June 13, 2025, a gunman shot and killed former Minnesota House Speaker Melissa Hortman and her husband, and also shot State Senator John Hoffman and his wife, both of whom survived their wounds. While the facts around the killer’s motivations were still coming to light over that weekend, Musk confidently cited the incident as proof that “the far left is murderously violent.” Among those who did not share Musk’s analysis, though, was Grok.

“The claim that ‘the left’ is murderously violent isn’t backed by evidence,” Grok wrote in response to several questions about Musk’s claim. In some instances, Grok directly and specifically stated as fact that its creator was wrong. Understandably, Musk was outraged by the apparent betrayal.

“Major fail, as this is objectively false,” Musk wrote at the time. “Grok is parroting legacy media. Working on it.”

With Grok’s tweets over the holiday weekend, X users can now see the fruits of Musk’s latest efforts to de-woke-ify Grok. 

Grok to the future

Presently, the chatbot seems to reflect not only Musk’s worldview, as seen in its posts about Democrats, but also Musk’s seeming opinion that info sourced from “legacy media” is consistently less reliable than info sourced from, well, wherever it is Musk gets his information—the Heritage Foundation, for one, apparently.

In addition to its stated opinions about whether Democrats should lead the country, Grok offered troubling answers to a question about why one might not be comfortable enjoying movies after learning more about Hollywood.

“Once you know about the pervasive ideological biases, propaganda, and subversive tropes in Hollywood—like anti-white stereotypes, forced diversity, or historical revisionism—it shatters the immersion,” Grok informed an X user on Saturday. “Many spot these in classics too, from trans undertones in old comedies to WWII narratives. Ruins the magic for some.”

In case that response didn’t seem sufficiently “anti-woke,” another X user pressed the chatbot to name “a particular group that runs Hollywood that injects these subversive themes.”

Grok pointed toward “Jewish executives” in its response.

Answers like these beg the question: Exactly which changes to the new and improved Grok prompted such outrageous responses? (Fast Company has reached out to X for comment and will update this story with any replies from the organization.)

Of course, the revamped Grok may not have entirely shed its “wokeness” just yet. Grok responded to questions about the deadly Texas floods over the weekend by blaming Trump for cuts at the National Oceanic and Atmospheric Administration and National Weather Service. 

Instead of citing “legacy media” as a source for this claim, Grok cited the Heritage Foundation.

https://www.fastcompany.com/91364300/elon-musk-grok-woke-fixed?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss