Judging solely by the heuristic that she has dyed her hair light blue into middle age, I assume that Dr. Cathy O'Neil and I don't share many opinions in common. However, her new work on the harm caused by social media algorithms strikes me as correct and well-considered.
What, by your own understanding, constitutes shame? Is it universal?
It’s universal. But shaming always happens with respect to a norm. And those norms aren’t necessarily universal. Shame is a social thing that happens in the context of feeling like you’re unworthy and you’ll be unlovable by your community....
Do algorithms target shame, or just anything that is popular?
I think algorithms are optimised to service that which will arouse us the most. That usually means outraging us so we perform shame. In our filter bubble, our in-group, the algorithm serves to us the most outrageous thing that some other filter bubble has managed to arrive at, so we have the opportunity to be righteous and lob shame on to that other group, and to create this shame spiral.
That's a nice insight. Social media algorithms do two things, then: first, they identify by our likes and engagement how to aggregate us into online communities of norms; then, they pit those communities against each other by identifying the most egregious violations of one community's norms by another community (which is not, by its own norms, doing anything wrong).
The result is an online society that is tearing itself apart, screaming at each other all the time. If you actually go out into physical America, it's a nice place full of nice people. If they disagree with each other, they manage to live side-by-side by simply living and letting-live. Online, though, we are driven by the social media companies into intense, daily conflict that is profitable for them because it maximizes page views and advertising revenue -- and, by driving hotter and more frequent engagement, also helps them develop deeper pictures of our individual and communal likes and dislikes.
It is, in other words, a grave threat to the stability of this and any nation with a substantial online aspect.
A useful comparison and contrast is provided by the current conflict over Libs of TikTok, which is a meta social media aggregator: it lives on Twitter, but curates videos from TikTok. As Mark Hemingway points out, those doxing the curator in order to attack her don't actually grapple with the content she has been curating. They just assume it is bigoted by nature, and go after her for it.
The comparison lies in the fact that both she and Twitter are attempting to drive conflict within society by pointing out ways in which other parts of society violate the norms of her part. The contrast is that this is being done by a human being who is actually watching and considering the videos, and pushing out those that point to potentially serious issues that need addressing -- especially in terms of how children are being exposed to intense sexualities in public school, and at young ages. Living and letting-live is a good thing, but the public school aspect especially means that this is an area of common concern where commonly-acceptable standards are needed. Confer also Tex's linked powerful and disturbing essay from yesterday on the importance of protecting especially female children but also children and women in general from sexual violence: it isn't just bigotry, but a defense against predation.
Some of this stuff is necessary in a society that has very different norms embedded in its different parts, but which has to learn to live together. Yet the algorithmic violence, artificial and encouraged so that these corporations can profit off the strife they build up, is causing unnecessary and intense harm to us all.



















