Just ahead of Hanukkah, the Jewish Federation of Broward County in Florida tried to purchase an ad on Facebook, a simple post calling attention to the problem of antisemitism as a part of a new nationwide campaign called “Shine a Light.”

But the Menlo Park-based social media giant’s automated system rejected the ad without explanation, leading federation officials to suspect the post was accidentally ensnared by a filter designed to block hate speech. 

“Unfortunately, Facebook inexplicably rejected our ads, presumably because they contained the words ‘hate’ and ‘anti-Semitism,’” wrote the federation’s board chair Alan Cohn and interim president and CEO Mark Freedman in a letter to the company on Tuesday. “This, we believe, is an unintended, but calamitous consequence of your effort to curb hate speech.”

If they are right, it wouldn’t be the first time such a thing has happened. From the moment Facebook banned Holocaust denial on its platforms last October, Jewish museums and other institutions doing education and outreach on antisemitism have reported trouble getting content published. 

When algorithms governing what can be published get it wrong, human reviewers are supposed to step in and distinguish hate from efforts to combat it. But many have found that getting their posts reviewed by a person and then approved can take months, if it happens at all.

In one notable recent instance reported on by J., Chabad rabbis across the country reported that their efforts to promote a four-part course about antisemitism on Facebook were blocked.

“It’s a real problem for us,” an official at the U.S. Holocaust Memorial and Museum told the Jewish Telegraphic Agency in June.

The letter from the Broward County federation, addressed to Facebook’s Mark Zuckerberg and Sheryl Sandberg, said efforts to restrict hate speech are appropriate but that Facebook’s current system “does as much harm as it does good.”

The federation asked that Facebook let its ad run and called on the company to amplify voices that combat hate.