What, specifically, is Facebook doing wrong today? How could they improve?
They rank content by the degree to which it captures users’ attentions and prompts sharing. This ranking incentive is fundamentally driven by their advertising business model, where attention from eyeballs and shares generates revenue.
Things that make us angry get shared more; numerous studies show this. An ad-driven social media service is this fundamentally driven to show us content that pisses us off.
> How could they improve?
By facing consequences for the violence they cause and profit from.
Would that satisfy your concern?
That's a way Facebook could be incentivized to improve, but the question was "how could they improve?".
I imagine in a lot of cases Facebook could correctly identify the sort of posts that people get angry enough about to lead to physical violence if it put anywhere near the level of effort into researching that problem that it does into maximizing its share of the user's attention. Such posts could then be given a low rank in Facebook's algorithm.
They make it easy to do bad things to a lot of people at once.
Expecting our communication channels to sort out the good from the bad is how we get to Chinese levels of censorship.
I also wasn't trying to argue that removing ads exclusively solves the problems.
The engagement objectives of the ad seller (Facebook, here) means that highly-divisive or inflammetory content, which generates pageviews, or increases time-on-site, may be promoted over more socially beneficial material.
I enjoy the thought of an alternate universe where the most common social network is a p2p federated open-source system without ads, just running on alternative firmware in people's WiFi routers. However, that universe would still have astroturfing, and it would likely be harder to detect.