3D-printed images of the logos of Facebook and parent company Meta Platforms are seen on a laptop keyboard in this illustration taken on November 2, 2021. REUTERS/Dado Ruvic

Since October 7, 2023, antisemitism has surged across the globe — not only through physical attacks or explicit rhetoric, but through something subtler: a coded language designed to evade detection while spreading hatred.

Contemporary antisemitism rarely presents itself with the overt brutality of the past. It no longer relies primarily on swastikas painted on walls, leaflets distributed on street corners, or incendiary speeches in public squares. It has learned something more sophisticated and perhaps more dangerous: communication through codes.

This is precisely the phenomenon documented in the guide to the Codes of Hate, produced by Brazilian criminal lawyer Lilia Frankenthal, founder of Aviva18, an organization dedicated to combating antisemitism and hate speech.

On social media, prejudice now travels through emojis, numbers, memes, and seemingly innocent symbols. A juice box. A rat. A watermelon. Numbers like “88,” “14/88,” or “23:16.” Triple parentheses around names. References to “globalists,” “financial elites,” or “world control.” To those outside this universe, these appear as disconnected noise.

To those who know the extremist grammar, they function as signals of recognition, intimidation, and belonging.

The problem, as the guide demonstrates with precision, is not the isolated symbol. It is how the symbol is being used. A juice box emoji in a conversation about food means nothing beyond a beverage. The same emoji repeatedly posted beneath Jewish content is something else entirely. The origin lies in an English pun: “juice” sounds like “Jews.” The phonetic similarity became a code and the code became an instrument of ridicule and dehumanization, reducing people to disposable objects.

The watermelon, widely recognized today as a symbol of Palestinian solidarity, can also be deployed in coordinated “visual spam” campaigns targeting Jewish users, where it ceases to function as political expression and instead becomes a tool of intimidation and harassment. Context is everything. A symbol has no fixed meaning outside the circumstances in which it is deployed. Other seemingly harmless emojis have also entered this lexicon: the brain, used to insinuate intellectual manipulation; the DNA strand, invoked to promote myths of genetic purity and racial superiority; and the combination of a globe and a money bag, which evokes the centuries-old conspiracy theory of Jewish control over global finance.

Even the Magen David, a sacred symbol of faith and identity, is frequently hijacked. When linked to dollar signs or messages of “hidden power,” it ceases to function as a religious icon and becomes a marker of racist stereotyping. This is the same logic as classic Nazi propaganda, dressed in the apparently harmless aesthetic of an everyday emoji — like the rat emoji.

It can simply be an animal. Or it can deliberately echo Nazi propaganda that compared Jews to vermin and plagues to justify persecution and extermination. The Nazi propaganda film Der ewige Jude (The Eternal Jew) remains one of the most brutal examples of this technique.

Dehumanization serves a precise function: no one begins by defending the extermination of people they fully recognize as human. First, those people must be transformed into something else. The emoji has merely modernized the method.

The number 88 is unambiguous to those who know the code: H is the eighth letter of the alphabet, and “88” means “HH” — an abbreviation for a Nazi salute. The number 14 refers to the “14 words,” a white supremacist slogan still used as an ideological marker. Together, they form “1488” — an explicit declaration of neo-Nazi alignment. Triple parentheses — (((name))) — function as digital identification markers, echoing historical practices of segregation: identify, separate, expose.

The Internet has simply accelerated a much older mechanism. Antisemitism has always known how to change clothes when explicit discourse became socially uncomfortable. In the Middle Ages, Jews were accused of poisoning wells and spreading disease. Then came the “international bankers,” the “rootless cosmopolitans,” the “controllers of the press.” Nazism refined the operation through bureaucratic euphemisms designed to administer horror: “final solution,” “resettlement,” “special treatment.” The words sounded neutral. The result was genocide.

Today, the logic endures. The format changes; the structure survives. And coded language offers something extremely convenient to those who use it: the permanent possibility of denial. “It was just a joke.” “It was just an emoji.” “It was merely political criticism.” “You’re overreacting.” The ambiguity is not accidental. It is a strategy.

Digital platforms were never designed to interpret historical subtext, irony, or extremist semiotics. Their moderation systems are trained to identify explicit slurs, not constantly evolving symbolic ecosystems. It is precisely in this blind spot that coded hatred thrives, giving modern antisemitism one of its greatest advantages online. Memes travel faster than manifestos. Irony dilutes accountability. Algorithms reward engagement regardless of whether what circulates is information or radicalization.

Combating contemporary antisemitism therefore requires abandoning naïve or purely literalist analysis. Not every symbol automatically carries discriminatory intent, and trivializing the concept weakens legitimate denunciations. But ignoring context is equally dangerous. The law, digital platforms, schools, businesses, and society at large must understand that modern hate speech operates through layers, references, and shared codes, not only through open declarations.

Several democracies have tightened their legislation against Nazi symbols, Holocaust denial, and antisemitic speech in recent years. Germany, France, Austria, Canada, Australia, and Brazil have all moved in this direction. Brazil offers an important example. Since the landmark Ellwanger ruling, the country’s Supreme Court has recognized antisemitism as a form of racism, an unbailable and imprescriptible offense. Yet even legal systems that criminalize antisemitic conduct face the same challenge: the law evolves far more slowly than the language of hate. Law is necessary. It is not sufficient.

What Frankenthal proposes goes beyond punishment.  She proposes literacy. And history suggests that without it, every piece of legislation arrives too late.

Digital radicalization rarely begins with explicit violence. It begins with normalization. First the meme. Then the irony. Then desensitization. Then dehumanization. History has already shown, with a clarity that requires no embellishment, where that road ends.

Recognizing these codes is not censorship, nor interpretive paranoia. It is civic literacy for the 21st century. Every authoritarian ideology depends on collective tolerance to grow. When no one reacts, the extremist understands that the environment is safe. The code works because it circulates without consequence, quietly infiltrating the social vocabulary.

Antisemitism has learned to speak in code. The challenge facing democratic societies is no longer simply to condemn hatred, but to recognize it before it hides behind irony, emojis, and plausible deniability. The most dangerous hate speech today is not always the one that shouts. It is often the one that smiles, posts an emoji, and insists it meant nothing.

Nira Broner Worcman is a Brazilian journalist, CEO of Art Presse, and author of A Sisyphean Task (translated from the Brazilian hors commerce title Enxugando Gelo), on media coverage of the war between Israel and terrorist groups.