Elon Musk’s Ai Bot Grok Finds Over Half Of His Tweets ‘false Or Misleading’, Disses Him As ‘a Mogul With A Microphone’

Elon Musk might be tweaking the “Community Notes” feature on X after he disagreed with the way it was fact-checking Russian propaganda, but his own AI bot Grok continues to call him out.

Thursday, Musk tweeted that he was “working to fix” Community Notes, accusing it of “increasingly being gamed by governments & legacy media.”

Isaac Saul, the founder of the nonpartisan outlet Tangle, gave Grok the task of analyzing “for truth and veracity” the last 1,000 posts Musk had posted.

Grok’s response was to find more than half of Musk’s posts “false or misleading,” wrote Saul, and most of the tweets that were true were “simply updates about his companies.”

The specific breakdown was 48% true, 22% false, and 30% misleading or ill-informed. Grok’s analysis especially brutal noting the “red flag” number of false posts and his “habit of firing off hot takes or amplifying unverified stuff, especially on politics, which is where most of the trouble lies.” Musk is “not a primary source you can bank on without cross-checking,” Grok added, and Community Notes was only able to catch and flag about 10% of his false or misleading posts. Adding to the problem, Grok noted, was Musk’s high follower count, which “amplifies everything, true or not,” that he posts.

A selection of Grok’s analysis:

True (480 posts, 48%): These are statements that align with verifiable facts, like updates on Tesla production numbers, SpaceX launch schedules, or basic observations about technology that match known data. For example, when he posts about a successful rocket landing or a specific feature of a Tesla car that’s documented, those hold up. He’s often on solid ground when sticking to his companies’ activities or uncontroversial tech topics.

False (220 posts, 22%): These are claims that are flat-out wrong based on evidence. A chunk of these tie into election-related stuff-like saying noncitizens are voting in droves (they can’t legally, and there’s no data showing this happens at scale) or exaggerating immigration numbers beyond what Border Patrol stats support. Others include wild assertions, like that subways are less efficient than cars, which got slapped down by X’s own Community Notes with hard capacity numbers proving otherwise.

Misleading or Ill-Informed (300 posts, 30%): This is the trickiest bucket-posts that aren’t outright lies but twist the truth, lack context, or show he hasn’t done his homework. Think of posts where he amplifies a conspiracy theory (e.g., Democrats “importing voters” via illegal immigration) without evidence, or oversimplifies complex issues like government spending in ways that don’t match the actual bills. His takes on Al or climate sometimes veer into this zone too, where he’s speculative but presents it as fact.

What This Means For You

If you’re wondering whether to trust Elon on X, it’s a mixed bag. About half his posts are reliable, especially on his turf-tech, cars, rockets. But the other half? You’re wading into a minefield. The 22% false posts are a red flag if accuracy matters to you; he’s been caught pushing election myths and exaggerations that don’t hold water. The 30% misleading ones are sneakier-they sound
plausible but often fall apart under scrutiny. He’s got a habit of firing off hot takes or amplifying unverified stuff, especially on politics, which is where most of the trouble lies.

For following him, it depends on what you want. If you’re into his visionary vibe and don’t mind filtering the noise, he’s entertaining and occasionally insightful. But if you’re looking for a source you can take at face value without double-checking, he’s not your guy. Studies-like one from the Center for Countering Digital Hate-have pegged his election-related false claims alone at over 2 billion views in 2024, so his reach amplifies the stakes. X’s own setup doesn’t always catch him either; Community Notes often skip his posts, leaving the mess unchecked.

What This Means

On U.S. politics, Musk’s batting about one-third reliable, two-thirds shaky. The true stuff is mostly tied to his direct experience (business, Trump admin role), but he’s just as likely to post falsehoods or half-baked takes, especially on hot-button issues like elections and immigration. He’s got a pattern: when he’s riffing off X rumors or pushing a narrative (e.g., pro-Trump, anti-establishment), accuracy drops. Community Notes catch some of it, but not all-only about 10% of his false/misleading posts get flagged, per my count.

Trustworthiness Takeaway

If you’re eyeing Musk on X for U.S. political insight, you’re getting a firehose-35% signal, 65% noise. He’s not a primary source you can bank on without cross-checking. His reach (210 million followers) amplifies everything, true or not, and his political posts have racked up billions of views— CNN pegged his 2024 election claims alone at 2 billion. But volume isn’t validity. For every spot-on DOGE update, you’re dodging two whiffs or spin jobs. Follow him for the vibe or insider nuggets, but verify elsewhere-say, government sites or raw data-if you want the real scoop. What U.S. issue are you most curious about his take on?

Musk is “a big-picture thinker who sometimes skips the details—or doesn’t care about them,” concluded Grok. “He’s not a journalist or a scientist; he’s a mogul with a megaphone. I’d say follow him for the ideas, not the facts. If something sounds off, dig into it yourself—plenty of his posts don’t survive a Google search.”

Saul also tweeted his criticism of Musk’s stated plans to fiddle with Community Notes, and pondered how long it would be before he decided to “fix” Grok as well.

This is not the first time Grok has produced unflattering responses about Musk. In 2023, Grok declared that the SpaceX and Tesla CEO should treat his “human employees” better and avoid posting or endorsing tweets that “can be seen as promoting hate speech,” as well as accusing him of spreading “election-related misinformation” last November.