An increasingly factious, mentally disrupted, and confused age needs to find a path to clarity and hope -- and it may be right in front of us, grounded in our own local neighborhoods
EDITOR'S NOTE: This post is part of a series of discussion topics for the upcoming 61st International Making Cities Livable (IMCL) conference in Cortona, Italy, October 29-November 1, 2024.
The psychologist Jonathan Haidt has reported deeply troubling news: the kids are not alright.
In his new book The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness, he describes the rise of what he calls "phone-based childhood," marked by the widespread adoption of smartphones and social media in the early 2010s. This shift coincided with significant increases in anxiety, depression, and self-harm among adolescents, particularly girls. Haidt notes that girls are more affected because they spend more time on social media platforms like Instagram and Pinterest, which emphasize appearance and social validation, leading to increased stress and anxiety.
This trend has coincided with the decline of what Haidt calls "play-based childhood," a trend that began in the 1980s due to increased (and largely irrational) parental fears of kidnapping, the decline of safe walkable and bikable neighborhoods, and the resulting decline in opportunities for unsupervised outdoor play. Haidt argues that free play is crucial for children's brain development and social skills, and its reduction has led to increased anxiety and difficulty handling normal childhood fears and risks.
Haidt describes the importance of "antifragile" traits, making it easier for children (and later adults) to cope with stress. The idea of antifragility, developed by statistician Nassim Nicholas Taleb, refers to systems that actually benefit from shocks and volatility, as opposed to those that are merely robust or resilient. Examples include muscles, immune systems, and ecologies. In each case, the systems grow stronger in the presence of stressors (exercise, pathogens, or periodic wildfires, say). When we over-protect these systems from stress -- preventing exercise, isolating from all germs, or snuffing out small manageable burns -- we actually make these systems weaker.
This, Haidt believes, is what has happened to a generation of children, as they have lost the daily "exercise" (not only physical, but also mental) of childhood play, with all its minor dangers and stresses. In its place, they now experience an artificial world of online interaction, at once overly protected and overly threatening. Caught between these extremes, children have no opportunity to exercise their coping skills in controlled doses. The result is that they become the opposite of antifragile.
Crucially, children naturally develop healthy coping traits when playing outdoors in their neighborhoods, interacting with others within the public and private spaces of the city. They learn how to explore, test boundaries, adapt to consequences, and form bonds with others, even as they experience occasional (controlled) dangers and hostilities from others too. Unlike in the online world, with its artificial, placeless forms of contact, their explorations within the networks of places around them allows them to build up practical skills and real-world judgments.
This is not the only problem with social media and the Internet, of course. The replacement of "play-based childhood" with "phone-based childhood," with all its disturbing consequences, parallels other troubling Internet-based phenomena: the rise of fake news, surging conspiracy theories, and increasing online bullying, by both children and adults. There are worrisome repercussions for political divisiveness, the decline of faith in institutions (and their integrity as a result), and the erosion of interpersonal trust, often translating into growing social isolation at the neighborhood level.
It's worth remembering that many people once thought the Internet would bring about a kind of golden age of knowledge, a leveling of expertise and a growth of capabilities. But too much knowledge, it turns out, is not knowledge -- it's noise.
Moreover, reliable knowledge has a crucial feature: it's deeply embedded in a relational network of other knowledge, verifications, test outcomes, and evidence. This network grows out of one's personal knowledge and judgments, and it extends into the more institutional forms of knowledge on which we all depend (say, weather forecasting, medical prescriptions, and so on). As this "deep net" structure grows, our knowledge builds up its confirmations and the confidence in its reliability. It is this reliability that enables us to fly in jets, cure many diseases, land on the Moon, or perform many other routine marvels.
By contrast, the Internet we experience lacks this deep-net structure. While "surfing" (i.e. only skimming its surface), we are regularly inundated by clickbait messages like "You won't believe the five stupid things you'll click on, number three will shock you!" (Says who? Where did this come from? Why am I being told this? Probably to sell you something, or to make you angry -- or to manipulate you in some other way.)
We suppose that we are in control, choosing what we view and where we go next. This is an illusion, of course: it is the algorithms that control us, expertly doing so in ways that maximize profits and other benefits for the manipulators. We are being fed a regular diet of very shallow-net structures indeed: titillating videos, enraging memes, provocative posts. We have no way of sourcing their claims, or checking citations, or reviewing evidence.
Contrast this "shallow net" structure with, say, the structure of Wikipedia. Although perhaps not perfect, it has become the go-to reference guide on myriad topics. It is certainly vastly more reliable than most Internet material, and evidence suggests it is remarkably accurate and useful on an impressive range of topics. That's in large part because it has a structure of required citations and hyperlinks -- the same "deep net" structure of knowledge that allows us to land on the Moon, or put up a telescope a million miles from Earth. If a reader believes a Wikipedia article is wrong, they can trace the source of the claim and view the evidence for themselves. If they have more correct information and can demonstrate that to the editors, they can even correct the article.
This "deep net" structure of knowledge illustrated by Wikipedia is also a feature of artificial intelligence. Its "neural networks" mimic the processes of the human brain, evolving remarkably intelligent outputs based on so-called "large language models," and other deep-net repositories of information. The technology is indeed powerful, and for that reason, dangerous. It is also enormously promising in a number of ways. The challenge for AI is the same as the challenge for the Internet three decades ago: how can we harness this structure to truly increase, and not decrease, human well-being? This will be a daunting and sobering challenge for the years ahead.
Interestingly and importantly, the structure of Wikipedia grew out of the structure of wiki, which in turn was developed by software engineer Ward Cunningham as a tool to share so-called "pattern languages of programming." Pattern languages, in turn, were developed by the architect Christopher Alexander, as a way of capturing the web-like interconnected structures of the best human environments. (Deep nets, it seems, have an important counterpart in the world of human environments.)
Patterns describe the essential relationships between the elements of a design, both within the design and outside of it. In both the built environment and in software, each pattern is hyperlinked to others and to citation sources, just as a Wikipedia page is hyperlinked to others, and to citations. This network of patterns and citations forms a deep-net structure, potentially vast depending on the depth of connections. Alexander intended pattern languages to be capable of capturing the more profound interconnected qualities of the best human places -- including the "deep nets" of evolved human traditions.
Ward Cunningham has said that wikis are, in fact little pattern languages, with each wiki page hyperlinked to others. In both cases, the deep-net structure is capable of capturing an essential aspect of the realm at hand. In the case of software, as in the case of the built environment, it is the web-like essence of a design problem, or how its structure of elements relate to each other and to other external elements.
The idea that pattern languages can capture the real deep-net structure of human places is certainly Alexander's great insight -- and it's all the more remarkable that this insight later became the basis of a highly useful innovation in the software world... and then yet another highly useful innovation in the world of knowledge and information.
The deep-net structure of human places also has a relationship to other biological structures. For example, ants and other social insects lay down information in their environments through chemical scent markers, enabling successive insects to build up highly complex and sophisticated structures. Examples include termite mounds that are capable of regulating their temperature and performing other functions. This phenomenon, known as stigmergy, likely plays a critical role in human environments too, as we form patterns of information that others build on successively. Perhaps a true "smart city" will need to take advantage of this phenomenon.
Moire fundamentally, the deep-net structure of human places, while often overlooked, has critical implications for human well-being. We can choose to enhance these web-like interconnected structures within our neighborhoods, fostering connectivity, enjoyment, adventure, exercise, and social contact. Or we can create neighborhoods that cut us off from one another, and leave us with little to do but retreat indoors to our screens.
For adults as well as children, this is an avoidable tragedy.
Below is a 20-minute summary of many of these ideas, in a February 2024 talk titled "Confronting the Age of AI: Wider Lessons for Settlement, Technology and Culture."
- Michael Mehaffy