The Death of History

The tools that once exposed and debunked Holocaust denial are powerless against AI and the algorithm. Is there a remedy?

As the algorithmic age takes Holocaust denial mainstream, historians must admit a bitter defeat. When the two of us met at Harvard 15 years ago, one was a professor teaching modern history, and the other was an undergraduate interested in the future of artificial intelligence. The elder of us had published a succession of books concerned with one of the most important and intractable questions of the 19th and 20th centuries: Why did the successful economic and social integration of the Jews after their emancipation or westward emigration produce such a vitriolic backlash?

I (Ferguson) had written a history of the Rothschild banking dynasty and a biography of Siegmund Warburg. I had studied the German hyperinflation of 1923 and its corrosive impact on the Weimar Republic. I had written books about the two world wars, one of which—The War of the World—offered a novel interpretation of the Holocaust as one (the worst) of a series of episodes of mass murder. I not only lectured on these issues at Harvard and other universities, but I also sought to reach the widest possible audience by engaging in journalism and broadcasting, producing more than 30 hours of documentary television for PBS, and Channel 4 or the BBC in Britain.

My animating goal was to ensure that citizens learned the lessons of history—including the vital lesson that antisemitic rhetoric can be translated into genocidal reality if it is adopted by a modern state. Yet today I see this hatred normalized, the Holocaust trivialized. Popular and influential conservatives such as Tucker Carlson and the Conservative Political Action Conference (CPAC) crowd now tolerate traffickers in Nazi revisionism. And the poison being peddled to young conservatives by radicalizers such as Nick Fuentes is even worse than traditional Holocaust denial. They don’t so much deny the Holocaust as mock it and celebrate it.

As AI permeates social networks, this kind of anti-history is spreading like a cancer. More than a generation after William F. Buckley exorcised it from American conservatism, antisemitism is back—not only on the radical right, but also on the radical left, where its purveyors cleverly blend it with anti-Zionism and anti-capitalism. As it is young voters who are most susceptible to these trends, the implications for the future of American politics are chilling.

The historian has to admit: I failed utterly. (I do not admit that lightly.) And so I join with the younger scholar of modern technology, Levin, to offer an explanation—and a remedy, if there is one.

The return of Holocaust denial was not something we foresaw. From the 1960s through the 2000s, it had been essentially an exercise in historical fraud, perpetrated by cranks. In dingy strip-mall offices and suburban basements, a handful of scummy think tanks and lone wing nuts churned out mimeographed journals and thick self-published books disputing the overwhelming historical evidence that the Nazis carried out an industrial-scale genocide that killed around six million people. These tracts were full of distortions and lies, but they often aped the form of respectable history in attempting to convince readers of factual claims about things like Waffen-SS pay stubs or the logistics of death-camp crematoria. This specificity opened them both to scholarly debunking and legal liability. And the whole thing was profoundly uncool—hardly anyone wanted to dive deep on the chemistry of Zyklon B gas with an unshowered slob handing out homemade literature from the trunk of his shabby car.

By contrast, the podcast boom of the late 2010s saw the rise of “anti-history”—a new brand of Holocaust denial that threw off the shackles of quotes and footnotes altogether in favor of freewheeling YouTube rants and edgy Rumble livestreams. Its standard-bearers were well-groomed and affable. They wooed viewers with the transgressive thrill of supposedly forbidden knowledge. Showcased by Carlson in late 2024 as “the best and most honest popular historian” in America, anti-historian Darryl Cooper jumped straight to explosive moral conclusions: “Churchill was the chief villain of the Second World War.” Although “millions of people ended up dead” in Adolf Hitler’s camps, Cooper said, the Nazis had simply misjudged the logistics of capturing so many people, so may have finished them off to spare them, humanely, the agonies of starvation.

Having dropped their bombshells, such anti-historians usually retreated whenever challenged by real scholars, vaguely conceding some Nazis atrocities and insisting that they were ordinary folks “just asking questions.”

The millennial and Gen-Z firebrands who followed them in the TikTok era have quickly dropped even this pretense. They mainly deny the Holocaust not by disputing historical events, but by brazenly doubting that the mass slaughter of Jews was wrong in the first place. Myron Gaines, a leading “men’s rights” influencer whose macho finance and dating advice has propelled him to over a quarter-billion views on YouTube alone, pulls no punches: “Yeah we like Hitler,” he posted to X in October last year. “No one gives a fuck what you woke jews think anymore. [Hitler] was a revolutionary leader and saved germany. The jews declared war on Germany first.”

On a 2025 episode of Gaines’s Fresh&Fit Podcast, conversation turned to the influence of Jews in America today. Gaines’s co-host, Walter Weekes, asked: “How do we take them down?” Guest Kadriyanna James chimed in: “We gotta kill the motherfuckers.” The original video was deleted, but the viral clips spread across social media.

Other haters with smaller platforms shrewdly amplify their virality with gimmicks that prompt even people repulsed by the message to watch and share onward. Paul Miller, a neo-Nazi who streams under the name GypsyCrusader, gets millions of views dressing as pop culture characters such as the Joker and Mario while doing Hitler salutes and threatening Jews. “I want to accelerate full-on race war now in my lifetime,” he vowed. “Let’s do it now and get it over with.”

Yet most young men have played enough Call of Duty to know that such openly bloodthirsty swastika-wearers can’t be the good guys. And that’s where Nick Fuentes comes in. A 27-year-old Rumble streamer who first earned far-right fame by attending the Charlottesville, Virginia, white supremacist rally in 2017, Fuentes presents himself as a history-savvy truth-teller. In contrast to competitors crouched over webcams in Red Bull–strewn bedrooms, he projects a retro talk-show vibe—suit and tie, tasteful set, warm lighting, high-end camera. Five nights a week, his million-strong legion of followers gathers to see Fuentes spend two to three hours reacting to the news and opining on pop-culture drama before engaging directly with fans via “superchats” they pay him handsomely to read on-air.

But when you ask one of them to explain Fuentes’s ideas, the answer is usually hopelessly vague. “You’ve just gotta watch for yourself,” they reply, attaching a video longer than Lawrence of Arabia.

This is a tell—a sign that the true attraction isn’t evidence or arguments, but instead the emotions Fuentes arouses. If you watch a three-hour lecture about history or science, you can summarize the key points to a friend, even if you can’t recall every detail. By contrast, there are no summarizable takeaways from performances like a Fuentes show. He gleefully spouts the N-word and unfurls slur-laced harangues against transsexuals, gays, Muslims, Indians, blacks, Jews, and women—language his white male listeners know would get them fired. Often, the offensiveness is unmoored from any ideology at all, like Fuentes’s recent declaration that child sex-trafficker Jeffrey Epstein was “cool as fuck.” The posture is that he’s freeing his audience—especially “involuntarily celibate” (incel) Zoomer men—from all the norms they resent.

With that pretext, Fuentes gives his viewers a fig leaf of cover to treat virulent hatred as a big joke they’re all in on. He praises Jim Crow segregation, admires Joseph Stalin, lauds the Taliban, and calls for an American dictatorship. But it’s all wrapped in ironic 4chan memes and intermixed with chat about Star Wars, Taylor Swift, and the Barbie movie. So Fuentes is only breaking one more taboo when he says: “Hitler was awesome,” “Hitler was right,” and “the Holocaust didn’t happen.”

Most young men have played enough Call of Duty to know that such openly bloodthirsty swastika-wearers can’t be the good guys. That’s where Nick Fuentes comes in.

The main point isn’t to convince people of that factual claim. It’s to make them reject the topic as unworthy of moral concern altogether, and to paint those who care about Nazi atrocities as cringey, out-of-touch scolds.

Last December, Piers Morgan invited Fuentes to his YouTube show and challenged him on his history of Holocaust denial. Fuentes appeared to recant his view on the number of Jews who were killed. “It’s at least six million,” he deadpanned. “But it could be a hundred times more than that.” Morgan totally missed the sarcasm. He played an old clip of his guest making jokes trivializing the Holocaust. “You think that’s funny?” Morgan asked indignantly, setting up what would ordinarily be a gotcha. Fuentes simply smirked back: “Why, too soon?”

The kids watching in the live chat reacted as if it was a heavyweight knockout. Morgan pressed on with a clip of British journalist Danny Finkelstein’s heartfelt testimony that Hitler “tried to starve my mum to death, did starve my grandmother to death.” Fuentes scoffed: “My generation, we’re just done with the pearl-clutching.” And he ridiculed Finkelstein, sing-songing in mocking imitation: “ ‘Me mum! Me mum! Me mum got killed by Hitler!’ ”

Highlights of the exchange flooded social media overnight. Stripped of context, the sound bites on TikTok and Instagram Reels conveyed zero historical argument. What millions saw was a young, brash, and charismatic outsider running rings around a broadcaster identified with the boomer establishment. Once users started watching, it was hard for them to look away. Fuentes dances on the cliff-edge between definitely serious and definitely kidding, insisting, “I don’t hate anybody. I don’t hate Jews,” even as he calls for American Jews to be rounded up and deported to the Middle East.

None of this would have been possible before social media—or, to be more accurate, algorithmic media. The algorithms are trained to maximize user engagement—metrics such as watch time, likes, comments, and shares. Platforms trying to sell ad space need stars who can attract eyeballs and make their owners spend hours arguing with each other below a video.

Although Fuentes himself has been banned from most top platforms for terms of service violations, companies seem happy to harvest revenue from clips of him, and from clips of people reacting to those clips (and so on, often several piles deep). Now add today’s AI into the mix. Since the original Facebook news feed, the AI used by the platforms has become vastly more sophisticated. It now systematically plucks such “rage bait” videos from obscurity and pushes them to precisely the users most likely to engage—on a scale, at a speed, and with a precision inconceivable 10 years ago.

The app’s interface may look the same, but show interest in a Fuentes clip, and quite quickly every other video is someone ranting that Jews (or blacks, or slutty women, or Democrats) are ruining America. The platform rewards brevity, though, and soon the algorithm promotes shorter content ever further removed from a coherent argument. Increasingly popular are 5- to 9-second shots of a pro-Nazi influencer such as “Dilano” or “Lawst,” wordlessly mugging into his front-facing camera, then poking his eye as a signal to “notice” Jewish control over the world. Also trending are nostalgic images of Gothic cathedrals or happy white people in 1950s suburbia, incongruously combined with blaring techno and slogans such as: “TAKE EVROPE BACK” or “THIS IS WHAT THE [juice box emojis, signifying Jews] TOOK FROM YOU.” For the health-conscious, some posts warn of “goyslop”—ultra-processed foods supposedly marketed by Jews to poison gentiles. Farther down the rabbit hole are caricatures of evil Jews or vintage Nazi art of Aryan soldiers, captioned with Holocaust denial references such as “271,000” (purportedly the real death toll) or “McGraw-Hill” (the idea that Jewish-owned textbook publishers fabricated the genocide). In the deepest bilges of the algorithm are calls to violence, often evading the platform’s content-moderation software with a peaceful alpine landscape and spoonerized incitement: “Jill kews.”

Increasingly, the creators of such content use AI-generated images to heighten the emotional impact of the message—innocuous scenes that turn into giant swastikas if you unfocus your eyes, or imaginary scenarios of angelic blond girls surrounded by brown-skinned rapists. Realistic deepfakes depict Jewish celebrities making inflammatory comments, or hordes of savages rampaging through a city. And, since 2024, AI has been used to resurrect the Führer himself. Viral videos now translate Hitler’s speeches into English using his own voice and delivery style—transforming what sounds to anglophone ears like cartoonish bombast into something far more compelling. Recently, as an experiment, we watched a few right-leaning TikToks. Six of the next 10 videos the algorithm served us were of Hitler oratory.

Even more alarming than the posts themselves is the impression that almost everybody in the comments section agrees with their sentiments.

But do they?

Click on a few accounts calling for all “Third Worlders” to be deported or trans people to be fed into wood chippers, and you’ll notice an eerie sameness. Hashtag-laden bios like “Patriotic American, army veteran, mom of 6. #MAGA #Christian.” Generic usernames like “AmericaDirst1488” or “Alex5932934713.” Blank profile pictures, or photos of a square-jawed man or a gorgeous woman that, if run through checker software, are flagged as probably AI-generated. In many cases, these accounts are unkempt juveniles somewhere in Eurasia, toiling for pennies in a grimy troll farm.

But a growing proportion are autonomous bots, whose persuasiveness is rapidly increasing as AI gets smarter. Yesterday’s bots could spam nose emojis (“this person is a Jew”) or GIFs of dancing Hasids (“the Jews are behind this”). Today’s bots can read comments and write emotionally resonant responses like, “I feel you. A Jew raped my sister, too.” Tomorrow’s bots will form ongoing relationships with users and target their personal anxieties.

This army of bots creates a powerful illusion of consensus. Especially for young people, seeing a large crowd affirming such messages legitimizes them, and stirs feelings of belonging for those who join. So, what starts with AI manipulation soon draws in all-too-real Americans—not only the incels, but also their friends and family members. Some of them post “the Austrian Painter [Hitler] was right” or “6MWE” (“six million wasn’t enough”) under their real names or from their business accounts, with smiling profile pictures of themselves hugging their spouses and children. Such behavior may begin under Holocaust denial videos, but it quickly spreads to the comments sections of mainstream conservative pages such as those of Tucker Carlson, Turning Point USA, and The Babylon Bee. Many MAGA influencers now find that antisemitic memes are among the most-liked comments on their videos. Some fight a sincere battle to delete such sewage as it gets posted. Some do not.

The results for the TikTok generation are alarming. Young Americans have become much more antisemitic than their elders, with 18 percent of 18- to 22-year-old voters saying the Jewish community has had a negative impact on the United States, compared with around 6 percent of those age 45 and older. Although both far-left and far-right radicals are much more likely to hold such views than the general population, the horseshoe is not symmetrical. Those ages 18–34 were asked whether they agreed with at least one of a series of antisemitic statements such as “Jews in the United States have too much power.” Among “extremely conservative” voters, 64 percent agreed with one or more, compared to just 34 percent of those who are “extremely liberal.” The Fuentes effect on this age bracket is clear—older, extremely conservative voters were not significantly more antisemitic than the overall electorate.

The Iran war brought these sentiments to the fore. As young MAGA voters rushed online to express their disillusionment with the war—within a month of its beginning, only 49 percent of Republicans 18–29 were in favor, versus 84 percent of those over 65—leading hate-mongers could pitch their Jews-cause-all-wars message to a massive new audience. This is already translating to offline politics. In Florida’s GOP gubernatorial primary, an endorsement from Fuentes has propelled far-right provocateur James Fishback—who rails against “goyslop” and said a black journalist “should be lynched”—into second place, with one month to go until the primary. Even among mainstream Republicans, a sizable minority of young D.C. staffers identify as pro-Fuentes “Groypers,” shaping the party’s rhetoric going into the crucial midterms.

Despite assurances that they take this problem seriously, the algorithmic media companies are losing control. They are simply not investing enough in the strong “immune-system” AI they would need to keep pace with the latest bots and their neo-Nazi creators. Instead, Meta, TikTok, and X still rely on antiquated content-moderation AI, and have all turned to crowdsourced “Community Notes” or “Footnotes” that let users flag false statements and vote on whether to append a fact-check. In this way, Elon Musk insisted, “if somebody tries to push a falsehood, like Holocaust denial, they can immediately be corrected.”

It’s not working. The Center for Countering Digital Hate estimated that, from February 2024 through January 2025, there were almost 680,000 antisemitic posts on X that likely violated the platform’s hate-speech rules, amassing 193 million views. The center identified the 100 most-viewed Holocaust denial posts and found that only two had a visible Community Note. Even when posts were flagged, they accrued about 78 percent of their total views before the note appeared. Censorship and bans have likewise largely failed. When TikTok started restricting juice box emojis, agile antisemites switched to kite emojis because “kite” rhymes with a derogatory term for a Jew. Decentralized hate networks such as “NazTok” are leaderless and resilient, so when the corporation nukes one eye-poker’s account, others emerge like Hydra heads to take their place.

We face, in short, the gravest challenge yet to the West’s commitment to free speech: an AI-powered assault on historical truth.

Of course, humans in America have a right to hold even the most odious views, and to share falsehoods peacefully with those who’ll listen. And we can still try to fight ordinary anti-history with the facts. But this is no longer the world in which Christopher Browning and Richard Evans could forensically expose the Holocaust denial distortions in David Irving’s books, ensuring that Irving lost his 2000 libel case against Deborah Lipstadt and Penguin Books. When Holocaust denial is beamed straight into the visual cortex as eight-second selfies of a teenager poking his eye, there is nothing to debunk, and no counterargument to make. In the time it takes to say, “On the contrary, as Raul Hilberg convincingly demonstrated in his magisterial treatise . . .” a stimulation-hungry Zoomer is on to the next video.

Indeed, the threat posed by the likes of Fuentes might be survivable if users were interacting only with other authentic humans. But they’re being inundated by a tsunami of AI slop and microtargeted rage bait on platforms whose algorithms reward “please stop mocking my grandmother’s murder :(” as “engagement.” No matter how loudly flesh-and-blood users shout, they are bound to be drowned out by a chorus of lifelike bots programmed to hijack the conversation. When they argued for free speech, neither John Milton, John Locke, nor John Stuart Mill could have foreseen a public sphere so polluted.

This will be a very hard problem to solve. AI is already smart enough to get through most CAPTCHAs—those “select all squares with traffic lights” tests that used to stump old-fashioned bots. And the tests can’t be made much harder without locking out many real humans, especially elderly or disabled users. (The challenge recalls an apocryphal quote attributed to a Yosemite National Park ranger, on the impossibility of designing bear-proof garbage bins: “There is considerable overlap between the intelligence of the smartest bears and the dumbest tourists.”) Once AI matches the general intelligence of humans, likely around the end of this decade, such capability tests will be altogether worthless.

We shall therefore need robust new proofs of humanness. One option is requiring users to authenticate themselves with a secure document, such as a passport or the UK’s proposed BritCard system. But this could destroy anonymity, enabling (and likely tempting) governments to track and persecute dissidents. More promising are decentralized technologies to authenticate real humans biometrically without revealing their names. For now, however, these remain quite flawed solutions.

As Pope Leo XIV argued in Magnifica humanitas, his landmark encyclical on AI, no top-down intervention can rescue us from this mess. We cannot engineer our way out of a deeply human truth: The algorithms reflect our own vices back to us—and magnify them. As long as we indulge our wrathful impulses, we’ll see more rage bait that stokes them. As long as we give in to our short attention spans, our news feeds will carpet-bomb our brains with dopamine. As long as we guzzle flattering lies, the truth will be muscled off our screens by propaganda.

This is a classic collective action problem. If you eat a Big Mac, it won’t make others unhealthy. But if you watch rage bait, you not only pollute your own information intake, but also tell the algorithm to spread that content to your friends and neighbors. So, just as we stigmatize littering, smoking indoors, or sneezing uncovered in a crowded elevator, we must learn to stigmatize the vices that befoul our shared digital environment. We must recover a shared sense of moral responsibility not to contribute to the conditions that let a Fuentes thrive. It’s a lesson that dates back to Cleon in ancient Athens. Ultimately, as Plato and Aristotle discovered, the only bulwark against demagogues is a people virtuous enough to resist their lure.

We fervently hope Americans still can be virtuous. Philosophers, artists, and spiritual leaders may yet call us back to the path of virtue. But it is up to them. In the face of an avalanche of AI-powered anti-history, the real historians are powerless. We thought we could prevail in a fair fight about what really happened in the past. We naively believed that, if we went to the right archives, and read the right documents, and summarized them with the right footnotes, and published our scholarly books and peer-reviewed articles, then the Holocaust deniers would be driven from the field.

What fools we were.

This article was first published in The Free Press on July 15, 2026