Reframing the “Social Media” Problem As an Attention Crisis
Every journalist has a “take” on Social Media. And most people are familiar with the battle lines now around “fake news”, “free speech”, etc., etc. Tens of millions have seen The Social Dilemma. (And if you haven’t, you should!)
Despite this, I’ve found that most people still do not understand what “social media” software actually does. Even sophisticated software and internet technologists wonder whether it’s really different than previous modes of communications. Consequently, most people aren’t understanding its true power and problems.
In this essay, I lay out these fundamental differences. I explain why I think “Censorship vs. Free Speech” is the wrong framing of the problem. And I will try to give readers a better framing of how to think about Attention, Trust, Intimacy, Norms, Virality, and Fame in an “online” environment.
But first, a simple thought experiment will help illustrate the self-created challenges of “social media”.
The ShoutBox
Imagine that your friends tell you about an amazing new music venue on the edge of town. You all go there one night, and pull into a packed parking lot, near a gigantic warehouse. You can hear a dull thumping of bass booming from the structure. Clearly there is an awesome party inside!
As you step through thick metal doors, you see bright disco lights and you are blasted by a wall of impossibly loud sound — literally every kind of sound you can possibly imagine. In any real-world venue, you’d have either one band playing on stage, or there’d be multiples stages where you could listen to the band of your choice.
But in ShoutBox, you can see multiple stages, each with a huge stack of speakers, and you can hear heavy guitar, drums, and even see DJs spinning techno in the distance. Huge speakers suspended from the ceiling are blasting multiple bass beats that you can feel in your chest.
It is so loud that you instantly get a headache, but thankfully a doorman hands you a set of headphones, similar to what you see everyone else wearing. As you put them on, you feel immediately relief — they totally block out all of the loud sound, except in the direction you’re looking. As you try to talk to your friends around you, you also realize that the headphones have a built-in megaphone. Only by massively amplifying your voice, can your friends possibly hear you over the noise cancellation of their headphones. But since everyone is talking through megaphones, it only makes the noise situation worse.
Every few seconds, your headphone audio is interrupted to play a brief advertising jingle.
While this may seem like a ridiculous design for a music venue, the reality is that our Social Media apps behave exactly like these imaginary “noise-cancelling megaphones”. They amplify the output of each individual, choosing broadcast by default. These apps don’t encourage individuals to explicitly address an audience. In normal human interactions, we don’t go around shouting at everyone with megaphones. But in the ShoutBox, this is the default.
These antisocial defaults are by design. Social Media incentivizes individuals to produce as much content as possible, so that others get inundated and must use some type of filter. By intermediating every conversation, they gain an enormous amount of power.
To see an example of this, try setting any social app to only display content from just your friends, sorted chronologically. The difference in user experience is shocking.
And yet, because they need the content to drive engagement, no app asks: what if we just turn down the volume? How much junk content is too much? When and how did we decide that “everyone publishes to everybody, including random strangers” is a reasonable model of communications? And most importantly, does this communication model really help people make better sense of the world? Does it really help them connect better?
Principles to Better Understand Social Media
1. Limitless Amplification Creates Attention Scarcity
Online communities technically don’t have a “free speech” problem. Anybody can set up a discussion forum or a mailing list, and have a more powerful communications fabric than the British Empire ever had.
Rather, for groups of people and society at large, what we actually have is an amplification problem.
Amplification, at the scale we’ve achieved in modern society, creates a bulk scarcity of attention. That is actually the hard limit that is driving a sense of “censorship” when e.g. Twitter, Facebook, Youtube decide to “de-platform” someone. We are now so used to an amplified media environment (ShoutBox), that the lack of amplification is equated to actual censorship.
Thus, the historical principle enshrined in “freedom of the press” has now morphed into “access to amplification”. But since all the routes from pixel to eyeball are owned by private carriers (Apple, Google, Facebook, Amazon, etc.), revocation of this access is total and absolute, and utterly controlled by private actors. Almost no corporate actor by itself has a total monopoly on broadcast and distribution, but taken together, the telecom+social media companies form a cartel whose control over attention is nearly absolute.
2. Society’s Synchronous Attention Is A Commons
In the discussion around social media, I’ve not seen anyone bring up the fact that a group’s attention becomes more scarce as it becomes more synchronized. In the business world, we all know that having too many meetings literally kills bandwidth to get anything done. But the same is true of any human organization, including the entirety of society.
We must view individual attention as a societal good; and we should see society-wide “joint attention” as, in fact, a Commons, like fresh air or clean water. But unlike air or water, synchronous attention is a manufactured scarcity.
Anyone who has ever been in a position of leadership knows how critically important it is for a group to “know a thing”, and to “know that everyone knows”. In a business, this could be the corporate mission or quarterly goals. It’s why we have regular status meetings to “get everyone on the same page”.
A society is no different. Prior to Internet, we relied on broadcast technologies to be a memetic metronome, and to set a cadence of sense-making and norm-reinforcement for a very large and diverse tribe. The decoherence of this joint-attention erodes foundational intersubjective beliefs, and destroys society by allowing individuals to “spin off” into a thousand diverse little subjectivities.
In the current environment of information warfare and accelerated fracturing into memetic tribes, how we apportion control over this Commons is the single most important question facing every country. This principle — that the sense-making environment should be a socially-governed Commons — is a new one. It is only apparent now, because tech has gotten so powerful that it can flood and clog the information environment at such massive scale.
It is on this principle that we can establish a modernized version of the Fairness Doctrine, which strives to define a healthy information diet for all citizens. Furthermore, by recognizing that synchronicity is a key aspect of memetic power, tech companies may be able to engineer better solutions to balance amplification while mitigating potential harm, by modeling the temporal availability and spread of a particular piece of content.
Every Wall Street trader knows that the momentum of a stock’s price is as important as the price itself. The impact to the social Attention Commons is no different: how quickly everyone around us begins to talk about a thing is, recursively, also a “thing” unto itself. In 10 years’ time, any government that still exists in a coherent form will have produced
3. Individuals Should Be Sovereign Over Their Attention
Attention is one of the most fundamental components of a conscious identity. When something hijacks or manipulates a person’s attention, it’s actively eroding their independent self-hood as a conscious individual.
The giving and the receiving of attention are both intimate acts. These acts are the very basis of all human communications, and by extension, human societies. Any tools or technologies that mediate this act need to be designed with great care, to artfully maintain the human elements around this mutual expression of vulnerability. In-person human interactions are extremely rich in context, but recording and broadcast technologies invariably force trade-offs. A year into COVID–19, video conferencing overload has made everyone painfully aware that there is no substitute for being able to see and experience our fellow humans in person.
Along these lines, Fred Rogers brilliantly articulated: “I think the greatest gift you can ever give is an honest receiving of what a person has to offer.” Yet no social media application even begins to contemplate how to help people situate themselves, so they can honestly receive what others have to offer. We have acclimated to giving unrequited intimacy to algorithms whose sole purpose is to furiously slap more pixels into our eyeballs, to keep us from putting our phones down.
The primary Internet concern for individuals isn’t “freedom of speech”. Anyone can still set up a website or send an email, and reach anyone else electronically. As Howard Rheingold observed in his 2012 book, Net Smart: How To Thrive Online, the most important thing that is actually threatened is an individual’s sovereignty over their own attention.
4. Every Conversation Is A Space, and Humans Need Spaces With Norms
Whenever two people exchange some signals (or even when a person reflects on their past, or writes something to their future selves), their minds intuitively develop the sense of a “virtual space”. For a real conversation to happen, each person must make space for the other.
As a conversation grows beyond just a few people, it becomes paramount to establish norms of the space. We all intuitively know this when we meet in person, whether it’s a CEO stepping into a boardroom, or someone approaching a stranger at a bar. We also know that a 2 or 3 people conversation is very different than a large dinner table with 8 guests. A room of hundreds, and a stadium of thousands, is more different still.
Online forums are the same. The social spaces on the early Internet consisted of mailing lists and Usenet forums that had explicit rules, norms, and moderators. And each user ultimately had the final say: if they really wanted to ignore a toxic individual, most applications supported the concept of a blocklist or “k-file”, to blanket mute someone across all channels.
Starting in the mid-2000s, as Social Media apps sucked more user attention into their walled gardens, their one-size-fits-all approach to User Interface and User Experience, caused many of the smaller communities to die off. Additionally, by pulling millions of people into the potential audience for every post or tweet or photo, these companies force everyone to treat every conversation as a potential performance to millions of total strangers.
But conversations should not be performances, unless they are explicitly staged as such. Conversations should not be battlegrounds. We should not smash people together in memetic Thunderdomes where only combat, and not growth, is possible. The design of modern Social Media platforms has done tremendous damage to human relationships.
5. Virality Is A Bug That Creates a Lottery of Fame
McLuhan famously declared that “the medium is the message”. But few seem to recognize that Social Media isn’t a “medium” in any traditional sense. Because it transmits and amplifies different content in variable ways, responding to the content itself, it is an active participant. If it was merely a constant, static amplification, then we could apply our traditional tools for assessing accountability and limiting harm. These include libel laws and the (now-repealed) Fairness Doctrine.
This variable amplification at the heart of “viral” social media isn’t an accident or a software bug. It is a necessary component of driving engagement for their business model. It creates a “Lottery of Fame” for every post, every image. Instead of everyone getting their 15 minutes of fame, social media apps seduce you with the possibility of getting 15 seconds of micro-fame: You could be an influencer, reaching millions, trending in the sidebar! This hijacks our primate-level desire for status. It’s especially toxic for young users, who become aware of status games among peers at a very early age.
Social media apps mine attention from captive masses, and mint it into social capital. We are training entire generations of youth to treat their phones as digital slot machines — “One-finger bandits” that use dynamic amplification to provide the possibility of “going viral” to keep people swiping. This “manufactured scarcity” in social media is actually a specific instance of the broader principle around network effects, brilliantly articulated in this essay/podcast by James Currier and Eugene Wei: Status Games: Engineering Scarcity in a World of Abundance.
6. Receivers Feed Back to the Transmitter and Create Oppressive Fame
When we think of communications technology, there tends to be an emphasis on the transmitter, or the impact of the outputs. But what about the flip side of broadcast, when an individual is the recipient of a lot of concentrated attention? And when this attention is presented not as an intimate offering to initiate a genuine relationship, but rather an imposition and a demand?
This concentration of social expectation creates an obligate identity, fully controlled by the mob audience, and which is deeply unsettling. In the small, it is exactly the oppressive mechanism behind “cyberbullying”. At scale, it is exactly the dark side of Fame, which celebrities from Cardi B to Justine Bateman have commented on.
This started with broadcast technology. As McLuhan noted, “electric media” is able to create social expectation faster than the target of those expectations can evolve their true selves. Their identities then become slave to the machine, unless they quit altogether.
If Hollywood multi-millionaires find themselves crushed and brutalized by the deeply inhumane psychological stress of fame, what are we doing to an entire generation of youth, who have been raised on micro-dosing fame?
(To be fair, Virality is a recent dynamic that can only manifest in modern “compressed time/compressed sensing” memetic environments. But even before that, any concentration of attention creates a susceptibility to a homogenous, synchronous transmission of a meme. For example, Orson Welles’ broadcast of War of the Worlds wasn’t “viral”, but it was a memetic hijacking of traditional broadcast.)
De-platforming Isn’t the Problem
With the de-platforming of Trump and some domestic terrorism promoters in the weeks after the January 6th attack, everybody and their dog felt compelled to write blogs about “tech companies are too powerful!”
I got fed up and wrote some angry tweets. My frustration was due to the fact that it should be blindingly obvious to everyone that the major tech platforms have been “too powerful” for a very, very long time now. It took the banning of Trump for folks to realize this? Here’s a simple syllogism: If de-platforming Trump deprived him of some great power, then surely giving him a platform in the first place must have bestowed him with great power? And anyone who can give power to an American president, must surely be even more powerful?
Everyone should have seen this, all along — especially tech journalists! For the last 15 years, Every. Single. Time. You. Swipe. Or. Scroll. The. Platforms. Exercise. Their. Power. Every time you move the mouse to hover over an image, the Javascript in your browser is reporting you to banks of supercomputers, running the advertising exchanges. Every time you scroll up or down, or scroll fast or slow, your app is reporting your behavior back to Facebook or Instagram or Tiktok, so their software can build a more complete emotional and psychological profile of you. These companies have been trafficking in aggregated human attention and human behavioral patterns for over a decade. Everyone in Silicon Valley knows this, and is complicit in building up this giant, dystopian complex of surveillance capitalism.
I am so, so relieved and grateful to see Tim Cook taking a stand for user privacy.
“At a moment of rampant disinformation and conspiracy theories juiced by algorithms, we can no longer turn a blind eye to a theory of technology that says all engagement is good engagement — the longer the better — and all with the goal of collecting as much data as possible,” Cook said.
Kudos to the Apple team for taking a clear step in the right direction.
Part 2 to come: Forget Decentralization: We Need a Generative, Humane Network
I hope that this essay helped people understand that we should be thinking in terms of attention and the collective Attention Commons of a group. I did not list all of the damage from social media; for that, please see the excellent Ledger of Harms from the Center for Humane Technology.
In Part 2 of this essay, I will make the case that this is not about “breaking up” some centralized monopoly, but rather we need to re-think the entire question of “what is the Internet for, and how do we make it support human development?” The modern Web (and what the Internet has devolved into) is quite broken at an architectural level. They were designed, decades ago, to network computers in a time of low bandwidth. But today, we need a new information system that serves the needs of the human network. It must reinforce trust and deliver serendipity in an era of high bandwidth and infinite memory.
Addendum: To the CEOs of the social media giants
Your core assets are your models of user interests/psychology, as well as the natural serendipity latent within the network of the human social graph. (Your UIs are not a core feature; Easy publishing on the web was mostly solved before social media.)
Please leverage these core assets into new product offerings that respect a user’s agency over their attention. Remember, your networks are not the only ways for people to discover each other. Alternatives are possible. The largest digital social networks remain email and SMS, and upstart competitors will bootstrap off your network to achieve overnight scale. You are now on the disruptee, not disruptor, side of the Innovator’s Dilemma.
It was once said that the internet interprets censorship as damage and routes around it. Users are slowly waking up to the fact that free-to-use, infinite feeds are predatory. They will come to view viral amplification as cancer, and learn to shun it from their human networks. The young parents of today are not hapless Boomers content to swipe and click on ads. Many of us are deeply, deeply motivated to manifest a different vision of the digital Commons for our children. We have seen a joyful, quirky, creative Web get paved over and turned into a crowded, concrete ghetto filled with hustlers and thieves.
As long as your business models involve aggregating and trafficking attention, I — and others senior technologists like me — will see it as our moral imperative to shine a targeting laser on this critical failure in the business models of circa-2000s Social Media companies, until some new disruptor launches a tactical nuke at it.
I’m happy for the Amazons, Walmarts, eBays of the world to discover new kinds of user-friendly marketplaces. I’d be thrilled for new models to destroy the old world of advertising & the attention economy that it demands. If your business models are chained to the old-world paradigm of advertising, your business will share its fate.
Thanks to Jason S, Dave K, Travis S, Hugo B, and Paul F. for helpful comments & suggestions on the draft of this.
