A new investigation conducted by The Guardian has revealed Facebook’s bizarrely inconsistent rules on violence, racism and hate speech.

The guidelines, published this weekend, aim to clarify what content is acceptable to post on the social media site. Known as the “Facebook Files”, they were supposedly introduced as a way to help overworked Facebook moderators determine what is actually offensive – facilitating their ability to make snap decisions and minimise time-wasting. 

Unfortunately, concerns have already been raised over the complex and irregular nature of these policies. For example, videos of abortions are allowed to be shared on the site, as long as there is “no nudity” involved. Likewise, all photos of animal cruelty are acceptable if shown with a warning, as is most non-sexual child abuse (as long as it is not sadistic or celebratory).

Even more weirdly, users are also able to livestream their self-harm attempts. “We don’t want to censor or punish people in distress who are attempting suicide,” read one apparent policy update. “Experts have told us what’s best for these people’s safety is to let them livestream as long as they are engaging with viewers.”

Given Facebook’s reputation for overzealous censoring – the site has taken down numerous pieces of classic art and photography for violating its guidelines – the leniency of these guidelines comes as a bit of a surprise. How can attempting suicide on camera be more acceptable than an iconic piece of war photography? Why can footage of abortion surgeries be shared but not a 19th-century painting of a vagina?

A couple of the other, more inconsistent guidelines can be viewed below:

Above, moderators are given examples of the sort of “aspirational/conditional statements” that are fine to be left unchecked. This includes a threat to “little girl’s” face, and a cute promise to cut someone’s tongue out. According to the guidelines, these sorts of posts are fine as they are “facetious and unserious”, and couldn’t realistically cause “real world harm.”

This is where things get a little more complicated. The “facetious and unserious” vibe carries through when people are threatening to snap women’s necks or kick ginger people, but the boundaries become less clear when the attention turns to Trump. As he is a head of state, and therefore in a protected person category, calling for someone to “shoot” him is just not permissible. 

Monika Bickert, ‎Facebook’s head of global policy management, has since defended the guideline’s inconsistencies. According to her, regulating the site’s two billion users made reaching a clear consensus very difficult.

“We have a really diverse global community and people are going to have very different ideas about what is OK to share,” she said in a statement. “No matter where you draw the line there are always going to be some grey areas. For instance, the line between satire and humour and inappropriate content is sometimes very grey. It is very difficult to decide whether some things belong on the site or not.”

She added: “We feel responsible to our community to keep them safe and we feel very accountable. It’s absolutely our responsibility to keep on top of it. It’s a company commitment. We will continue to invest in proactively keeping the site safe, but we also want to empower people to report to us any content that breaches our standards.”