In her new book, “The Shame Machine,” the author and mathematician Cathy O’Neil (also the author of the wonderful “Weapons of Math Destruction”) discusses the “shame industry,” its creators and proponents, and what can be done about it. O’Neil argues that shame, once used to promote adherence to community values, has morphed into a destructive, demoralizing, profit-mongering industry. She weaves together personal anecdotes and scientific evidence to support her arguments against the shame industry and advocates for reclaiming shame as a tool for societal change. Overall, this book presents interesting and helpful food for thought on shame and its proper place in our world, although those interested in AI’s role in this industry may be disappointed by the paucity of material on this subject.
The book is divided into three sections – industrial shame, networked shame, and healthy shame – in which O’Neil explores different parts of shame culture. The section on industrial shame dives into four topics steeped in shame: weight loss, drugs, poverty, and beauty. Here, O’Neil argues that shame is (harmfully) propagated in these areas when those struggling are led to believe that “if you just tried harder, you wouldn’t be in this position.” She provides numerous anecdotes (and some scientific backing) for how this shaming attitude fails to bring people into a better place and is often exploited for revenue-generation.
Networked shame, the most relevant topic for the AI & Faith community, explores the role of social media news feeds in the propagation of shame. O’Neil notes that, although historically shame has been used to enforce common values in small communities, recommendation algorithms in newsfeeds perpetuate shame indiscriminately to large audiences. Recommendation algorithms promote the most engaged content, which often can be shaming due to the strong reactions such content provokes. Since shaming content is emotional, people are tempted to pile on, creating a huge online shame-storm targeting individuals who violate a real or perceived cultural norm. This, she argues, is the root of cancel culture.
I think we all can relate to the temptation to participate in public shaming on social media. I occasionally find myself mentally composing tweets about things that annoy or disappoint me and envisioning how others would probably agree with me (and how nice that would feel!). Fortunately, I lost my Twitter password, so I can’t follow through. However, O’Neil provides a forceful exposition of how unhelpful, unproductive, and cruel this sort of online shaming can be, especially when directed against people with little power. Her thoughtful and persuasive illumination of the problem of online shaming has encouraged me to avoid mentally writing those tweets at all and have more empathy towards those with opposing viewpoints.
Beyond large-scale online shaming, O’Neil argues that networked shame leads shamed individuals to retreat to corners of the internet where their actions are viewed as correct. As these subcommunities grow more isolated, they create a false sense of moral rectitude (since people with different opinions inherently are not in the subcommunity) and further perpetuate shaming behavior. As an example, she discusses how the specification of personal pronouns is commonplace in some circles, but that when people outside these circles encounter this value, they are confused because their community has not espoused this value. This leads to shaming on both sides, as each side is shocked that the other violates a perceived moral standard.
O’Neil’s proposed counterbalance to the shame industry is “punching up,” using shame where appropriate to bring people in power in line. She highlights several recent movements, such as #MeToo, where public shaming has brought real societal change. However, this “punching up” must be deeply nuanced, as this kind of move is only productive when directed against people with power to make changes, rather than everyday people who could be drowned by the mob. She couples this with a call to avoid online discourse that encourages shame (e.g. “just don’t share that shaming video”).
So, what lessons can we as the AI & Faith community take from this book? Speaking to the faith side first, this book is a call to recognize the power of shame when it is grounded in a set of common values, which presents two opportunities for people of faith. First, people of faith are inherently people of values, and our divided world could use some common values. We have the opportunity and obligation to promote our values of divine love and human dignity to the world, to unify people otherwise divided. Second, as people who value human dignity, we have the opportunity (and responsibility) to speak against shaming, to avoid perpetuating shame machines and to challenge others to do the same. From the AI side, O’Neil’s specific discussion of networked shame as it relates to newsfeeds is also a call to action for faith-based AI practitioners. Now that some harmful side-effects of these algorithms are well-documented, these practitioners can lead (and should) the charge to mitigate these algorithms’ harms and enhance their positives.
Overall, this book is thought-provoking and attitude-checking. It encouraged me to consider more carefully how I interact with people with whom I disagree, both personally and in online settings. It broadened my understanding of shame itself and its place in our world. Again, although the AI-specific content of this book is limited, “The Shame Machine” is another great work from Cathy O’Neil.
