According to The New York Times, the new policy is largely a response to the role Facebook played in the violent attacks on Muslims that took place in Sri Lanka, Myanmar and India. In Sri Lanka, for instance, riots broke out after rumors went around on the platform, accusing Muslims of poisoning food given or sold to Buddhists. Facebook was also accused of helping promote violence against Rohingya Muslims in Myanmar by allowing misinformation to circulate on its website.
The company will reportedly use its image recognition tech to spot status updates that violate the new rule. CNBC says it will also work with local and international organizations to help it spot and verify fake news. See, while it plans to take down posts that encourage harm, it will only de-emphasize fake news that don't incite violence in people's Feeds. That's where those partner groups come in.
As for what kind of posts don't deserve to be banned, Facebook chief Mark Zuckerberg gave Holocaust denial as an example in his interview with Recode. That obviously caused an uproar, prompting him to clarify that while he "personally find[s] Holocaust denial deeply offensive," Facebook's goal is "not to prevent anyone from saying something untrue -- but to stop fake news and misinformation spreading across [its] services."