Off the top of my head, I’m not aware of any riots or protests that directly led to fatalities. Cases that result in fatalities usually did not originate as a large demonstration or riot and conversely most large-scale demonstrations/riots are often under just enough police control (e.g. separation from counter-protestors) to prevent fatalities from happening although injuries, throwing stones and similar occurances do happen more frequently.
Furthermore, the requirement of ‘caused by social media’ is difficult to prove. At best, the organisation of a demonstration (or subsequent riot) is facilitated by social media but as large scale demonstrations or riots are not a new observation (indeed, they date back to at least 1789 to allude to just one of the most prominent in history) so they certainly happened prior to Facebook/Twitter/messenger apps and the like.
If we relax the conditions a little bit, somewhat famously in Germany the (originally weekly) Pegida demonstrations began in 2014 with a closed Facebook group of the founder Lutz Bachmann. In a more recent example, the Chemitz protests, which led to a number of injuries, were also most likely co-organised via social media, likely on both sides. To give an example from the other side of the political spectrum: the far-left protests and riots during the G20 summit in Hamburg in all likelyhood had a social media component to their organisation. Again, no fatalities but over 2000 crimes committed according to the police reports.
It bears mentioning that at least part of the fear is supported by research: in 2018, Vosoughi, Roy and Aral analysed the spread of a number of news stories across Twitter between 2006 and 2017 and found that
Falsehood diffused significantly farther, faster, deeper, and more broadly than the truth in all categories of information, and [that] the effects were more pronounced for false political news than for false news about terrorism, natural disasters, science, urban legends, or financial information.
(Source: S. Vosoughi, D. Rob, S. Aral, Science 2018, 359, 1146–1151. DOI: 10.1126/science.aap9559.)
Their work considered practically every rumour that spread a type of ‘news’ (‘We also purposefully adopt a broad definition of the term news. Rather than defining what constitutes news on the basis of the institutional source of the assertions in a story, we refer to any asserted claim made on Twitter as news’) and they determined whether news was considered to be true or false by checking six independent fact-checking organisations which agreed on levels between 95 % and 98 %.
Thus, ‘false stories [which] inspired fear, disgust, and surprise in replies’ travel faster on Twitter and have a greater potential to spark protests.