Does Facebook Want You to See Antisemitism?

I worked in product development for decades. Here's why hate isn't a bug in Meta's business.

This is a multi-part analysis of how social media helps spread hate. In this part, we will look at mechanisms that increase the virality of hate and why Meta allows it to proliferate.

---

I’ve worked in tech for my entire adult life. The antisemitic content on Facebook has been growing and when I try to report it, I invariably get back a statement that there was no violation found.

This holds up no matter how vile the content is.

So I decided to dig into the why.

Meta’s Community Standard statement:
We believe that people use their voice and connect more freely when they don’t feel attacked on the basis of who they are. That is why we don’t allow hateful conduct on Facebook, Instagram, or Threads. [1]

And yet, no matter how hard I try to avoid antisemitic content, almost 100% of the Jewish content on my feed has it.

Let’s look at an example from JewBelong, a U.S. based group that fights antisemitism using meme-like posts and billboards. They posted something calling out Holocaust denial.

The original post

Instead of elevating thoughtful replies or corrections, Meta ranked as ‘most relevant’ the comments that denied the Holocaust, celebrated it, wished for it to return, or suggested the victims did something to deserve it.

Most relevant comments

The comments also suggested that living Holocaust victims are ‘too old to exist’, inverted the history into a Jewish crime, turned the genocide into a punchline, and 2 people openly wished for a second one.

All of this was under the default filter, ‘most relevant’.

---

If you’d like to support my writing, you can buy me a coffee ☕ . No pressure💛

---

The Engine: Meta’s Ranking Algorithm

In Meta’s documentation, relevance isn’t determined by the content of the comment so much as how much engagement it gets. Every time someone replies to a comment or reacts to it, Meta determines the comment has higher relevancy.

When you see Most relevant above the comments on a post, it means that comments you see are ranked, and you're more likely to see high-quality comments that are relevant to you. [2]

This is why arguing with people on Facebook, or adding a reaction to comments that make you sad or angry, actually boost the relevancy algorithm.

The system does not distinguish between people who engage to agree, or because they’re horrified. The only thing that counts is engagement.

So, arguing with antisemitic content or hitting the angry emoji is counterproductive. It feeds the algorithm and will increase the comment or post’s relevance, pushing it out to more people.

The Tactics: How Bad Actors Exploit the Engine

The Passive Mechanism: 🤣 Haha as a Weapon

The ‘haha’ reaction was designed as a light way to respond to something funny. But it’s more often used to express derision and cruelty. [3]

It bypasses content moderation since it’s not words. On public posts, authors can limit who is permitted to comment, but has no ability to control who can haha their post. The only option is to block the person and that doesn’t erase the haha count.

I’ve seen the laugh emoji used on news about synagogue attacks, Jews being stabbed, hostages in Gaza, and Holocaust remembrance. And when Jewish pain is mocked, Meta reads it as an engagement signal and sends the post or comments out to more people.

The Active Mechanisms

Coded Language

Antisemites have developed many ways to get around content moderation.

Mustache man or ‘the Austrian painter’ are used to reference Hitler without naming him. Mustache man was either ‘right’ or should have ‘finished the job’.

Coded examples:

• “Juice” or an emoji of a juice box - stands in for Jews. 🧃

• $7,000 - used to indicate the amount a person was paid to say nice things about Jews or Israel

• “Globalists”, “the elite”, and “Epstein class” often function as coded references to Jewish control and pedophilia.

• “Epstein class” has an extra kick by intimating that Jews are the wealthy elite who control our politicians and also positions Jews as pedophiles.

• “Me mum was a shoe in the holly”, ”Me dad was a lampshade” are used as a stronger derision than haha on posts about Jews.

The meme below doesn’t name Jews and is designed to escape algorithmic detection. It calls on classic antisemitic tropes about ‘parasites’ and ‘hosts’ suggesting historical expulsions were justified. It distills old hatred into a bite size, easy to crosspost, meme.

Sometimes imagery doesn’t have to name Jews at all but can do the work with a swarthy, hook-nosed man.

Weaponized Shitposting

Shitposting is the creation and dissemination of aggressive or absurd memes designed to mock, derail, provoke, and get reactions. And it’s designed with deniability built in.

Below, James Gunderson uses ‘juice’ to reference Jews and makes a play on words on a short about Jewish American Heritage Month with a reference to concentration camps.

Because it’s wordplay, Gunderson can use total deniability if confronted.

This one uses juice to reference Jews and has a play on words “concentrated” to reference concentration camps. It’s right under ‘most relevant’.

Bot Networks

In addition to bad humans, automated bots play a huge role in spreading hate and misinformation. While some estimates put bots at 20-30% of social media accounts [4] there’s no way to find the exact statistics as social media companies don’t want this information out there. (It would impact advertising revenue.)

These aren’t just guys living in their mom’s basement making bots. Much of the bot infrastructure is funded and organized by state actors in places like Russia, China, and Iran with the intent to inflame existing social fractures, turn us against each other, and / or influence elections. At times, the rhetoric leads to real-life violence.

Bots are what may make hateful positions look mainstream.

Similarly to reporting hate speech, reporting bots generally leads to Facebook suggesting you block the person.

---

The Business Model: Why Meta Isn’t Interested in Fixing This

In early 2025, Meta removed fact-checking from its platform (similar to X) to prioritize ‘free expression’.

Meta’s platforms are built to be places where people can express themselves freely. That can be messy. On platforms where billions of people can have a voice, all the good, bad and ugly is on display. But that’s free expression. [5]

The same article indicates that past moderation had gone too far and harmless content was being censored. This policy statement is in direct contradiction with their Community Standards statement on hate speech:

We believe that people use their voice and connect more freely when they don’t feel attacked on the basis of who they are. That is why we don’t allow hateful conduct on Facebook, Instagram, or Threads. [6]

Rage is good for business

Meta’s product is attention. The platform makes money by getting people to scroll and interact so that it can show them more ads. Content that provokes emotion, especially outrage, is especially good at holding attention.

We don’t need to speculate about Meta’s motivations or practices. In 2021, former Facebook product manager, Frances Haugen leaked thousands of documents showing the company knew anger was one of the most reliable engagement drivers. On 60 Minutes, she said:

“Facebook makes more money when you consume more content. People enjoy engaging with things that elicit an emotional reaction. And the more anger that they get exposed to, the more they interact and the more they consume.” [7]

She also noted that Facebook put money over public interest:

“The thing I saw at Facebook over and over again was there were conflicts of interest between what was good for the public and what was good for Facebook. And Facebook, over and over again, chose to optimize for its own interests, like making more money.” [8]

Haugen estimated that Meta only takes action against 3-5% of hate and ~0.6% of ‘violence and incitement’. [9]

References

[1] Meta Community Standards

[2] What Most Relevant Means on a Facebook Page Post

[3] Laughing at death: Facebook, the ‘haha’ reaction, and death coverage on local US news pages

[4] Digital distortion: How bots are deepening America’s divide

[5] Meta Position Statement: More Speech and Fewer Mistakes

[6] Meta’s policy on Hateful Conduct

[7] Facebook whistleblower says company incentivizes “angry, polarizing, divisive content”

[8] Whistleblower: Facebook is misleading the public on progress against hate

[9] Facebook encourages hate speech for profit, says whistleblower