I'm afraid the concern is warranted.
TV wasn't terribly disruptive because it was kinda like the next logical progression of radio, which had long since been a sort of established, semi-institutionalized way of disseminating information. Not everyone could broadcast, the airwaves were regulated. Advertising revenue was also a sort of check on the quality of content as well. If your programming sucked, people would watch something else, and advertisers would stop paying the channel. The concentration of social media into just a handful of tech giants has led to a much more captive audience, I guess. Not only that, but platforms figured out that emotional manipulation is the best driver of engagement, and built it into the platform to maximize ad revenue. It doesn't really matter whether or not the content driving engagement is constructive, or even true. Rush Limbaugh figured out the same thing in like the early 90's on his radio show. In hindsight, I dunno why there weren't more Rush Limbaughs, but there was still no way to algorithmically build in emotional manipulation for radio and TV, at least, not like is possible with social media. The feedback loop is almost instantaneous for social media, whereas TV needs to react to Nielsen ratings, for example. I think the comparison made to the printing press in the article was pretty apt, though.
I really do think social media is leading to social destabilization worldwide. The way I would go about semi-proving it would be to develop two "system state vectors" for each country; One would be a measure of social well-being, using metrics like quality of life, economic growth, wealth inequality, artistic output, scientific literacy, etc. (some of these are difficult to quantify, and up for debate), and the other state vector would quantify how prolific social media usage is in a particular country, using web traffic data and self-reported statistics (obviously, a good assumption right off the bat is that social media use is constantly increasing everywhere, but the timing and rates will vary with how "developed" a country is). I wouldn't be surprised if the vast majority of the time the two state vectors correlated proportionally. Still, not definitive causation, of course.