Human nature is really shapeable — and we (as a former product manager in this space) designed algorithms and product experiences to maximize certain outcomes that benefit social network bottom lines. Regardless of how good or bad human nature is, the algorithms prize engagement however they can get it — I personally believe it surfaces a certain side of humanity, and pushes people in a certain direction.
If you're a news editor, you actually have a choice to sensationalize all day because it maximizes short-term profits. After bad things happened (see yellow journalism, the newspaper circulation wars) — newspapers did cut back a fair bit. (We can debate whether that was the risk of regulation, or their own respect for the role they had)
This clearly isn't the social networks' fault alone, but their product choices do partially influence on how we all behave. On HN, you hear a lot of anger at journalists for being so click-baity, and the sad part is that every incentive many journalists face encourages this (from user behavior to profit)
There are tons of examples how bad incentives can warp how people behave, far beyond "humanity is good" or "humanity is bad". For example, see Hannah Arendt's Banality of Evil or the Stanford Prison Experiment. More regularly, see architecture, urban planning, or product management in the Valley, as disciplines that think deeply about the outcomes they want to encourage.
A more accurate statement I would say is: humanity has certain predilections, and design choices can influence these dramatically.
(And now I will shill my media literacy for engineers guide — contributions welcome: https://github.com/nemild/hack-an-engineer )
My take away is that, after ten years, we can say that Facebook and Twitter have a negative impact on contemporary society.
What are the evidences for "People are generally good"?