On November 3, 2025, Elon Musk unleashed Grokipedia, his AI-powered encyclopedia, into the world. Musk, who’s never met a problem he didn’t think could be solved by throwing AI at it and then tweeting about it relentlessly, described Grokipedia as “better than Wikipedia”—which, apparently, for him and his right-wing fanbase, simply means “not woke.” Forget decades of academic scrutiny, open editorship, and community oversight: now, “truth” comes prepackaged from the richest guy in the room, delivered with a shrug and a snark.

Grokipedia arrived with Musk’s typically oversized bravado. He promised “the truth, the whole truth, and nothing but the truth.” Unfortunately, what users—and several high-profile scholars—found instead were factual errors, blatant Wikipedia plagiarism, editorializing, and disturbing right-wing dog whistles. Wikipedia often gets accused of having a liberal bias—mostly because reality, much to the chagrin of certain politically-charged billionaires, tends to have one too. So how did Musk fix it? Simple: replace the collaborative, transparent process with a proprietary black box, then pepper it with your personal Twitter takes.

Within days, it was obvious Grokipedia was less about advancing knowledge and more about advancing Musk’s pet narratives. It’s the academic equivalent of telling your teacher that your AI did your homework, and when they find out it just copied Wikipedia but replaced the citations with angry Twitter memes, you brag about “disrupting” education. Sorry, but if you’re searching for “the sum of human knowledge,” Grokipedia is the last place you should look.

Distinguished historian Sir Richard Evans, who knows more about the Third Reich than Musk knows about not posting, checked his Grokipedia entry and discovered… everything was wrong. His degrees and illustrious career path? Fictitious. Grokipedia’s AI had “hoovered up” a mountain of garbage data from the internet and uncritically spit it back out, unfiltered, giving a chatroom troll’s opinion the same weight as decades of peer-reviewed scholarship. It’s less “the wisdom of the crowd” and more “the fever dreams of Reddit, 4chan, and Elon’s mentions.”

Evans’s experience wasn’t a fluke. Entries on everything from Nazi war criminals to Marxist intellectuals were riddled with errors, distortions, and—crucially—right-wing revisionism. Even facts that have been publicly, repeatedly debunked kept popping up, not because they were true, but because Grokipedia’s AI can’t tell a conspiracy theory from established consensus if both are phrased with equal confidence in the training data.

David Larsson Heidenblad, deputy director at Sweden’s Lund Centre, drew a sharp line between how academia and Silicon Valley think about knowledge. In techbro-world, flailing and failing forward is a “feature, not a bug.” You screw up, you SoftBank your way out, iterate, then declare yourself a “genius” for learning nothing useful. In academia—where errors can fuel far-right ideology, Holocaust denial, or disinformation—this is a recipe not for innovation, but for disaster.

It doesn’t take a PhD to spot Grokipedia’s glaring right-wing bias. Musk calls Wikipedia “Wokepedia” (congrats, you’re extremely online), and Grokipedia is hellbent on doing the opposite—by which I mean parroting right-wing dogma at the expense of, you know, facts. Whole chunks of content were lifted from Wikipedia, but the places where it diverges are especially revealing.

Take the entry for Britain First—a far-right organization and literal hate group by court conviction. The Muskopedia? Oh, it’s just a “patriotic political party.” Their leader, a serial bigot and racist, is described in glowing, sanitized terms. The Russian invasion of Ukraine? Grokipedia treats Kremlin propaganda as legitimate sourcing, going so far as to echo Russian claims of “denazification.” The “Great Replacement” conspiracy theory? Given serious (and dangerous) consideration as having “empirical underpinnings.” In short, if it’s been condemned as xenophobic, racist, or outright fascist by actual experts—or, say, the courts—there’s a strong chance Grokipedia will “both sides” it, or worse, validate it.

Meanwhile, events like the January 6th insurrection are downplayed as mere “riots,” not the attempted violent coup, led and incited by MAGA and Trumpists, that it very obviously was. For anyone who actually cares about democracy, history, or, hell, basic decency, this should be alarming. Grokipedia tries so hard to reject “the left-wing narrative” that it’s blindly assigning credibility to literal fascists, authoritarians, and Putin’s propaganda mouthpieces. That’s not “balance.” That’s dangerous.

At heart, the whole Grokipedia debacle isn’t just amusing—it’s existentially worrying. For centuries, we’ve wrestled with who gets to define “knowledge” and how it’s recorded. Encyclopedias started as top-down, Eurocentric silos, then Britannica, and finally, Wikipedia—a crowd-driven, sometimes messy, but transparent and democratic experiment.

Grokipedia is the very opposite. It puts a billionaire technocrat and his preferred AI model (trained, we can safely assume, on a slurry of half-remembered Wikipedia entries, partisan news, unfiltered social media, and Musk’s own late-night Reddit scrolling) at the helm of what millions of people may blindly trust as fact. Musk is, by any rational standard, a questionable arbiter of “the whole truth.” When you can’t even tell how content is selected, edited, or moderated—because the algorithm itself is proprietary and secret—you’re not advancing transparency. You’re erasing it.

This amped-up smoke-and-mirrors act is not a bug, it’s a business model. There’s no community to challenge bad entries, no effective way to hold Grokipedia accountable. If Musk or his team want to tweak facts, push a narrative, or favor certain politicians (gee, wonder which side?), there’s no check and balance. In an era when conspiracy theories and fascist rhetoric spread at viral rates, this is a recipe for disaster.

The Wikimedia Foundation’s response to Grokipedia basically boils down to: “Yeah, no.” Wikipedia’s strength has always been its messy, public collaboration. There are loud debates, mistakes are made, but corrections come fast, and bias is actively policed. You can literally track every edit, every argument, and every user’s contribution. Problems? Sure. But sunshine has always been the best disinfectant.

In contrast, Grokipedia has already shown it will go out of its way to “balance” things for the right—or to make controversial hate-mongers and convicted criminals sound like upstanding “patriots.” It doesn’t even require human curation to boost bad-faith actors. Just sprinkle some AI, borrow from unreliable sources, and let the algorithm—overseen by a guy who has a meltdown if you say “cisgender” on his social media platform—sort it all out. I’m not feeling reassured.

If you care about truth, history, or even just not embarrassing yourself in front of actual experts, avoid relying on Grokipedia for your “research.” The fact that an AI can now spin up alternate histories and cloak them in a veneer of neutrality should worry anyone who cares about, well, civilization. This is especially true if you belong to a marginalized or targeted group—because Grokipedia is already showing a dangerous willingness to validate conspiracies against you.

So, let’s recap. Grokipedia is:

• Plagiarizing Wikipedia, then spinning the results with a distinct right-wing and Musk-flavored bias
• Spewing factual errors and legitimizing far-right talking points, some of them literally hate speech or propaganda
• Abandoning the last two decades’ worth of transparent, community-driven knowledge creation in favor of algorithmic wizardry managed by a single billionaire
• Happy to raise the profile of bigotry, conspiracy, and pseudo-scholarship

Critics aren’t being alarmist—they’re sounding the fire bell before this thing catches on. If you believe, foolishly, that data is objective just because it fell out of an AI, consider who designed the AI, who chose the training data, and who ultimately decides what gets shown or censored. Hint: It’s not you, and it damn sure isn’t the “objective” truth. The iron law of platforms is: bias is built in. It just shifts from the messy crowd to the single autocrat behind the curtain.

Grokipedia is a warning sign. Let’s not sleepwalk into a future where history is crowdsourced in boardrooms and billionaire penthouses, insulated from criticism and accountability. Next time someone tells you to “trust the AI,” ask who fed the beast—and who gets to profit from its ignorance.