Analysis filed Oct 20, 2026.

Record no. HR-0006
Logged: Oct 20, 2026
Platform and place: Facebook groups and pages in New York, New Jersey, Florida, Arizona and New Mexico
What: a month of hostile comments about Jews, Muslims and their institutions, posted under real names
Status: mostly speech below any legal threshold; one item (a posted map of synagogue addresses) logged as targeting
Names: no commenter is named under our standards

The people threatening Jews, Muslims and their neighbors online are not hiding anymore. They post with a profile photo, a hometown, an employer, a church and pictures of the grandkids. They are not eggs. They are the guy from the hardware store.

Facebook's whole theory of accountability was the real name. Make people post as themselves and they'll behave. For the past month The Hate Record has been reading local Facebook groups and pages across half a dozen states, and that theory is finished. The names are right there. The threats are right under them.

A car crashes outside a residential synagogue in New Jersey while prayers are underway. The local news page frames it around the political deal that approved the synagogue. The top comment jokes that the worshippers were probably all scrambling for a parking spot. Hundreds of reactions. Some of them are laughing faces.

A crash post in Rockland County draws a pile-on of coded jokes about a "driving school" and a county renamed for wrecks. No slur appears. Everyone in the thread knows exactly who the joke is about. That is the point of a code.

A page that bills itself as being about "equal treatment under the law" posts a directory of synagogue addresses with homes circled in red, plus photos of a private house and a crowd outside it at night. It gets shared into a neighborhood forum. That is not commentary. That is a map.

A new Holocaust education center in Phoenix opens and the comments under its announcements turn nasty within days. A planned Islamic center in New Mexico, not yet built, draws calls to desecrate the site before a shovel touches dirt. A new Jewish neighborhood near Lake Okeechobee becomes the target of an online campaign sustained enough that state troopers and the county sheriff drove out in person to hear residents' safety concerns.

And in one Facebook group about a small New York town, a man posting about Jews has a profile that lists the car dealership where he works. He did not hide it. He did not think he needed to.

Here's the inconvenient part. The ADL counted 6,274 antisemitic incidents in 2025, a 33 percent drop from 2024. Still the third-highest year since it started counting in 1979. But physical assaults hit a record 203, and the ADL's audit excludes most online incidents because there are too many to count.

CAIR, on the other side of the ledger, logged 8,683 anti-Muslim complaints in 2025, the highest since it began publishing in 1996. Minnesota's chapter alone nearly doubled.

So the incidents that get counted went down, and the stuff nobody counts went everywhere. The comments section is the uncounted country.

In January 2025, Meta moved most hate enforcement from proactive detection to waiting for users to report it. The results came out under oath this summer at Australia's royal commission on antisemitism.

Facebook took action on 5.8 million pieces of hateful content in late 2024. By July through September 2025, that was 1.2 million. Instagram fell from 7.4 million to 2 million. Roughly a 79 percent drop on Facebook. Meta's policy director agreed the number was in the ballpark.

Meta's position is that relying on user reports has not led to more antisemitism on its platforms. The same executive told the commission that Meta has no standalone antisemitism policy and does not track hateful conduct by the group it targets. None of the major platforms could tell the commission, from their own data, whether antisemitism was rising or falling.

You cannot say hate didn't rise when you don't count who it hits. That's not a finding. That's a blindfold with a press release.

The users noticed even if Meta didn't. In a survey of about 7,000 Facebook, Instagram and Threads users run by advocacy groups including GLAAD and UltraViolet, 72 percent said they see more hate aimed at protected groups, and more than a quarter said they had been targeted directly. Meta declined to comment on that survey.

Everyone assumes anonymity is the accelerant. The research says the opposite. A University of Zurich team led by Katja Rost analyzed 532,197 comments on 1,612 online petitions and found that people posting under their real names were more aggressive than anonymous ones, and that full-name haters were actually more common than anonymous trolls.

Their explanation is the whole story in three lines. The haters don't think they need to hide. A real name makes them more credible and helps them rally the people in their own network. And they assume they will almost never be held accountable.

The name isn't a risk. The name is the point. In a local Facebook group, the audience isn't the internet. It's the neighbors. The likes come from people you'll see at the gas station, the school board meeting, the Little League game. Signing your name is how you prove you mean it, and how you find out who agrees.

Every one of these threads starts with a trigger: a crash, a zoning vote, a new building, a rumor. The comment section then does what ANIA has watched it do again and again. Grievance first. Then dehumanizing language. Then unverified claims. Then "someone needs to do something." Then addresses. The circled synagogue directory is the last step on that list, not the first.

The tone from the top matters too. CAIR's 2026 report argues that rhetoric and policy from public officials helped drive last year's record anti-Muslim complaints. The ADL found 45 percent of antisemitic incidents in 2025 were tied to Israel or Zionism. The comment sections we read were mostly about something more local: land, buildings, traffic, and "them."

Hate this year doesn't carry a party registration. It carries a comments section.

Most of this is legal. Ugly, coordinated, cowardly and legal. Since Counterman v. Colorado in 2023, prosecuting a "true threat" requires showing the speaker at least recklessly disregarded that it would be taken as a threat. Police cannot arrest a man for a parking-spot joke under a synagogue crash.

The platform waits for a report. The police wait for a crime. The commenter waits for the next post. Nobody's job covers the space in between, and that is exactly where all of this lives. It is the space The Hate Record exists to cover.

Report it. Meta built its entire system around user reports. That makes your report the only lever Meta left on the table. Use it.

Preserve it. Screenshot the full page with the URL, the time and the member count before it gets deleted. Threads vanish the moment someone gets nervous.

Call police when it crosses the line. A posted address, a photo of a home, a named date, a plan to "show up." That's no longer an opinion. That's logistics.

And Meta could turn the tools back on. Its own executive told the royal commission the company already has the AI to detect antisemitic content at scale. It chose to wait.

Facebook asked for real names so people would behave. They gave their names. They just never thought anyone would read them.

We're reading.

Method and disclosure: examples come from ANIA's read-only monitoring of public Facebook groups and pages in New York, New Jersey, Florida, Arizona and New Mexico from early to late October 2026, logged at the thread level; no commenter is named, and ANIA keeps no files on people for speech alone. These examples are not a representative sample. National figures come from the ADL's 2025 audit (which excludes most online incidents), CAIR's 2026 civil rights report (which counts complaints, not verified incidents), a self-selected advocacy-group user survey, testimony before Australia's royal commission, and Rost, Stahel and Frey's 2016 study in PLoS ONE. Meta was asked for comment in early October on New Jersey groups described here. The publisher, Pesach Lattin, is an Orthodox Jew and has no financial relationship with Meta or any platform named. ANIA is a privately held, independent intelligence firm, not a government agency. Corrections and tips: [email protected] or 505 932 9060.