Turns out Facebook will let you focus advertising at people with some horrible worldviews. An investigation by ProPublica published Thursday revealed that the company’s advertising platform allows people to send ads specifically to people who list topics like “Jew hater” or “How to burn Jews” among their interests. Not only that, but it allowed ProPublica journalists to pay to promote posts to such groups.

When ProPublica pointed out the option to Facebook, the social network removed the categories. It also said that they had been created algorithmically, presumably by scraping user profiles and doing nothing to filter out offensive results. But Slate has since tried a similar approach and found that it could promote posts to people who listed “Nazi Elementary School” in their education history or “Ku Klux Klan” as an employer.

If only this were a new problem at Facebook. ProPublica has shown in the past that the company has allowed advertisers to exclude users by race. And its failure to consistently police content is well known: earlier this year, an investigation by the BBC showed that sexualized images of children lingered on Facebook pages.

Mark Zuckerberg has repeatedly argued that Facebook has a very fine line to tread between policing offensive content and acting as a censor. He also thinks that AI will solve these kinds of problems. But AI is still a long way from helping. Until it does, Facebook will have to ensure that its regular old humans do a better job of screening the social network for hate.