The advertisements were purchased through its self-service tool, which allows buyers to independently purchase and target ads, often without human interaction on Facebook's side of the transaction.
In response, Facebook's product management director Rob Leathern issued a statement to CBS News Thursday evening saying the company doesn't allow hate speech and that "there are times" when content sometimes violates its standards and that the company has "more work to do".
Facebook generated the categories automatically with the results of what users typed into their education and employment fields. Once ProPublica informed Facebook of these categories, it took them down.
Facebook has also come under fire for allowing fake accounts to place adverts that may have sought to influence the United States presidential election.
In its report, ProPublica noted that the objectionable ad categories were very small.
"We've removed the associated targeting fields in question".
France discusses increased pressure on North Korea with Trump, Abe
They say usually negotiators take time quietly behind the scenes to come up with these new sanctions. Pyongyang claims it tested a hydrogen bomb capable of sitting atop a ballistic missile.
Reports surfaced Thursday that Facebook's ad targeting algorithm was surfacing these inappropriate categories to advertisers. After recent white supremacist violence in Charlottesville, the CEO said Facebook would remove violent threats, and the company subsequently removed deleted certain neo-Nazi and white nationalist accounts that it had previously permitted.
After being contacted by ProPublica the groups were removed by Facebook.
This is a big deal because it could shape the styles of content created for Facebook Watch, the new original programming hub its launched where publishers earn 55% of ad revenue. Facebook spokesman Joe Osborne told ProPublica, "We have looked at the use of these audiences and campaigns and it's not common or widespread".
Here's how it worked: Advertisers target campaigns by specifying the groups of people they are interested in reaching. It said the entire process was approved within 15 minutes. In many cases, those categories were suggested by Facebook's automated system when typing words like "Jew" or "Hitler".
Considering all these elements, Facebook might soon have to draw a clearer line between not allowing hate speech on and not discriminating groups of people regardless of their thoughts and beliefs.