s08e18: QAnon looks like an alternate reality game

... and here's the subtitle saying "Here's How I Know..."

0.0 Context setting

It’s Wednesday, 5th August 2020.

This is a Very Special Episode of Things That Have Caught My Attention. Normally I write a combination of one or two longer-form observations about technology and what it’s like to be human, along with a few shorter links. This episode will be devoted to one topic.

The below is an adaptation of a thread of tweets I wrote on July 10, in response to a thread of my brother’s.

The ideas in these threads are not new—we are far, far from the first people to notice how the online behavior meshes with conspiracy theorists— but what I think my brother and I both bring is the perspective of having directly designed and run alternate reality games and their communities.

Since starting the below essay, life managed to intrude in various ways (I always wonder why we say this; as if there’s some conscious experience upon which life cannot intrude, which I’ll write about in the next subscriber-only newsletter episode).

My brother has also written an essay on his blog, What ARGS Can Teach Us About QAnon, as well as a follow-up interview with Charlie Warzel in the New York Times; Is QAnon the Most Dangerous Conspiracy Theory of the 21st Century?

Before you read what I and others have written, though, I encourage you to read work such as Mols Sauter’s The Illicit Aura of Information (2017) and The Apophenic Machine (2017). I wasn’t aware of Sauter’s work when I wrote my off-the-cuff Twitter thoughts; am grateful for having them brought to my attention. Sauter also graciously reminded me of the importance of citation and building on networks of previous work, not least of which because it helps move us forward and beyond working on thoughts again and again on our own.

I’m reminded that in the nearly 20 years since my involvement in alternate reality games we are fortunate to have had the practices of digital anthropology, digital sociology and science and technology studies develop. But these fields have had to struggle for funding, for legitimacy and for recognition, not least of which because they have traditionally been the subject of sexism in both academia and wilder culture. I’m but one person, not academically trained and merely curious with a personal interest, and it is clear to me that as a society, we might be much further in how we work with culture and society online if only we had treated these subjects with better care and attention.

We still have time.

1.0 QAnon looks like an alternate reality game

I used to be a game designer, making alternate reality games. Fans call them ARGs, and they’re played by hundreds, thousands, tens of thousands of people around the world who get hooked by a story, a challenge, puzzles and characters. But mainly, their players get hooked by the puzzles. 

I got involved in these games nearly 20 years ago, through a marketing campaign for the Steven Spielberg film A.I. This game, The Beast (so named because an early content production audit auspiciously resulted in a to-do list 666 items long), told the story of Evan Chan, who died under most suspicious circumstances. We needed to figure out who’d done it, and why. 

Set in the future of 2142, The Beast was a maze of websites, clues, emails, phone calls, faxes (it was 2001, after all) and live events teeming with actors. It was such a big deal in marketing and interactive entertainment that year that it ended up being one of Time’s ideas of the decade, and would later be called the first truly successful alternate reality game by The Atlantic. 

The trailer and poster for Spielberg’s A.I. included a credit for “Jeanine Salla” as a “Sentient Machine Therapist”. In credits that include directors and producers and actors, a Sentient Machine Therapist sticks out like a sore thumb, and so Googling Jeanine’s name in 2001 led you to the game’s first website. 

A secret you could discover, all by yourself or, more likely, after a tip from another fan.

(You can read a wonderful summary of The Beast in Cloudmaker days: a memoir of the A.I. game , by my friend Jay Bushman, available in the ACM’s digital library, as part of the book Well Played 2.0.)

Around six thousand people all over the world played The Beast game intensively. Obsessively, even. I know, because I and five other friends, and my brother, were community moderators who helped organize players’ efforts, manning the mailing lists and IRC channels.

The out-of-the-ordinary credit in A.I.’s trailer and poster would end up being called a rabbit hole by players and designers: once you fell down it, the point was to keep you falling deeper and deeper.

As game designers, our goal is to keep our players engaged and having fun. Sometimes, these goals end up getting switched around, or worse, perverted: you can definitely tell the difference between being engaged with something and having fun.

In an ARG, being engaged would mean solving the next puzzle or clue and unlocking a new piece of story, Alice in Wonderland-style. (ARGs love their literary references. The Beast had a character called, of course, the Rational Hatter).  

Nothing is new, of course. We’d learn we weren’t the first to tell stories in this way, as epistolary fictions through written evidence instead of narrative. Even the name of the genre, alternate reality game, would egg us on to tell a story through as many media as possible to flesh that alternate reality out, to make it feel real. We’d use letters (printed out! On paper!), emails, website contact forms, phone calls, and more. And we’d make the puzzles you’d need to solve to make sense of it all hard

The key bet that the designers of The Beast had made was to design for the curiosity not of an individual, but the collective curiosity of thousands. To design for the fact that the internet made it easier for people to do things together. They bet that that with thousands of players exploring their mystery, we’d figure out a way of coordinating our efforts. So they made their puzzles difficult: a sort of liberal arts meets technology extravaganza, a Voltron of triva, puzzle-solving and pattern-matching ability. Codes hidden in the Declaration of Independence (imagine being the one person who realizes a strange image superimposed on a graphic is exactly the same as forming part of John Hancock’s signature), codes hidden in medieval lute tablature, locations encoded in hidden latitudes and longitudes and more. 

No one person could possibly be expected to have all of this esoteric knowledge to hand: the only way to win, to solve the mystery, to find out who killed Evan Chan and why, would be to work together. 

The Beast went on for a few weeks. It got so complicated, so sprawling, that I started keeping a webpage — a trail — that acted as a written account of what we learned when, what puzzles we still needed to solve, what loose ends needed investigating or closing off. Where my document was an active to-do list, with increasingly esoteric and challenging puzzles crossed off, my brother wrote a guide: a passive, lean-back narrative retelling of the story, a sort of proto-Television Without Pity recap of the entire experience. For many, this experience of following on with the drama would be much more accessible than the down-in-the-weeds frenzy of speculation and puzzle solving.

In other words, to play this game, to get to the end, our community started making those walls of photos and post-its, notes joined by red string. A trope-y detective show crazy wall. We needed to make sense of everything. We didn’t put this kind of stuff in a Wiki because — remember — it was 2001, and people weren’t really using wikis like that back then. 

There’s a lot of pattern recognition going on, when you play games like this. There can be clues everywhere, so you look everywhere. Sometimes you get a hit. A lot of times you get a miss.

We’d make fun of this, playing The Beast: players would post their ideas with the tag SPEC or even WILD SPEC for particularly outlandish episodes of pattern-matching and red-string threading. If you posted an idea someone had already advanced, there was a chance you might get TROUTed, an in-joke for the players who hung out on IRC.

But for The Beast, and every ARG that followed, there was something important to remember: you’re playing a game that’s designed to be solved.

Right from the beginning, as community moderators and players, we’d talk about the experience of apophenia, the drive our brains have to make connections, to find patterns, to see faces in places

The use of apophenia quickly became a key part of the genre of ARGs: the compulsion for pattern finding ramped up a million-fold, explicitly designed in. And, unfortunately, variably reinforced, like the most addictive casino slot machines and algorithmic social feeds. Because just like in real life, in these games, not every theoretical connection would be true. Not every imagined connection would pay off. Some of them would, and those pay-offs would be designed to be exciting and dramatic. You say designing for addiction, we’d say designing for engaging narrative and playful, social experiences

But some people wouldn’t remember we were playing a game. They’d use tools to look up information about a website’s owner and then call that person up or, in some cases, go to their house, a sort of 2001-era doxing. 

When this happened in 2001, my fellow moderators and I would put a stop to that behavior whenever it happened. “This isn’t part of the game,” we’d say. Other players would look for security vulnerabilities in the game’s web servers and hack their way in. We’d say the same thing: “We’re playing a game. We’re supposed to be able to solve this. You’re peeking behind the curtain. We don’t need to cheat.”

But the nature of these games reinforced that behavior. In ARGs, some addresses to physical locations, hidden in images, emails and behind codes to break, were real and in-play. At the end of the day, if you really want to, you can turn anything in to a location if you want to massage numbers enough into the requisite GPS coordinates.

Traveling to the right locations would yield sweet progression of the story, narratives and yet more puzzles to solve in exciting ways: USB sticks, maps, audio recordings, more clues. More cards to put up on the wall. More connections to make. 

Even if you lived hundreds or thousands of miles away, the power of the group was that there would inevitably be someone who lived close enough and was motivated enough to drive, bus or travel by any means necessary to track down that clue. There was drama in watching the chase. Every single person playing was in the part of the guy in the chair, the trope skewered in 2017’s Spiderman Homecoming, providing technical support to the superheroes. And next time, you could be the superhero, too.  

Crowdsourcing, as a way of describing how groups of people come together to contribute ideas online was only coined in 2006 to describe emerging behavior. 

There’s been a lot more ARGs in the last 20 years. The ones I have the best memories of were for the TV show Lost; the Cloverfield series of films; Halo 2’s ilovebees; Nine Inch Nail’s Year Zero album release; and Potato Sack for Valve Software’s Portal 2. Cicada 3301, a set of online puzzles is thought of as an alternate reality game, and the inspiration for an episode of Jonathan Nolan’s Person of Interest. Unsurprisingly, many of the techniques in puzzle-solving ARGs ended up being used to recruit for state intelligence services.

In the months after The Beast, when I’d be working on these games myself and designing them, I’d learn about Richard Bartle’s taxonomy of player types, a 1996 paper based on research on people who play multiplayer online games.

Bartle suggested that there were four archetypes players would fall into: killers, achievers, socializers and explorers. His research would also be one of the elements feeding in to the early 2000s trend of gamification, of explicitly awarding points and scores and leading to concepts like Microsoft’s Xbox Live Achievements. Although Bartle had based his research in multiplayer online role-playing games, I quickly saw that we could classify every single ARG player into one of his four archetypes, too:

The killers were the ones who wanted to break things, or break other players. They were the trolls. 

The achievers were the ones also who just wanted to win, and sometimes winning would include cheating. 

Socializers were there for the chat and the company, or a more passive, tv-watching type experience. 

Lastly, the explorers were the ones interested in the story, in solving puzzles, actively feeling and finding out what would happen next. 

I’d refer to Bartle’s framework again and again over the next 20 years throughout my career, applying it not just to games but using it to understand and think about online behavior in general.

A few weeks ago, my brother brought up again a theory that’s been around for a while: that QAnon was like the ARGs we’d made and played together. He thought that QAnon was popular partly because doing the time-consuming research understand and contribute to the “QAnon community”, to be QAnon, is enjoyable because it’s active, not passive, like watching TV. This made a lot of sense to me, and it’s part of what’s scaring me about QAnon, too. 

Because if you look at QAnon through a lens of game design, it starts to look a lot like behavior I and my fellow game designers have seen amongst ARG players over the last 20 years. Only, clearly, a lot worse. Every single QAnon behavior I’ve seen feels like it’s at least an order of magnitude more intense than ARG player behavior—but uncontrolled, undirected, and unconstrained.

From the outside looking at how people take part in QAnon, there’s a lot of similarities: being a part of QAnon involves doing a lot of independent research. You can imagine the onboarding experience in terms of being exposed to some new phrases, Googling those phrases (which are specifically coded enough to lead to certain websites, and certain information). Finding something out, doing that independent research will give you a dopamine hit. You’ve discovered something, all by yourself. You’ve achieved something. You get to tell your friends about what you’ve discovered because now you know a secret that other people don’t. You’ve done something smart. 

We saw this in the games we designed. Players love to be the first person to do something. They love even more to tell everyone else about it. It’s like Crossfit. 

I brought up Bartle’s four player types earlier, and I see them represented amongst QAnon adherents, too:

QAnon’s socializers are meme-makers, and their success creates achievement and community standing. 

QAnon’s achievers are those who find connections that further the conspiracy, the ones who join the red string together on that board. They play for local fame: to be the first to find the connection. Their achievements are ripe to share and provide that socialization kick. Their local fame is quantified and made instantaneous through retweets, favorites and Facebook reshares. 

QAnon’s explorers? They’re the connection finders, too. But in ways different to ARGs, QAnon’s explorers get to create new material, too. In most of the designed ARGs with a defined story or stories, players only have limited ability to contribute to the world or plot. Explorers in the world of QAnon get to create new evidence. 

And lastly, QAnon has killers. The killers here are, well... charged with murder for actually, horrifically, killing people. And, as in the games world, they’re griefing (trolling and making life horrible for) people who aren’t doing QAnon properly and, crucially, trolling the people who aren’t even doing QAnon in the first place. Like the socializer meme-makers, the killers are spreading QAnon. Unlike most ARGs, QAnon has an in-game sense of enemies: people who don’t even have to be playing the game or aware the game exists, to be considered opponents.

When I was writing about this on Twitter, someone pointed out that a big difference between QAnon and ARGs was that, well… aren’t ARGs just a game? And wouldn’t QAnon provide an even bigger rush because it isn’t a game? I don’t think this quite tracks. Well-designed games, just like any other well-constructed media are able to lull you into a sense of disbelief. People who play ARGs want to believe it’s not a game partly because when done right, we’ve designed them so compellingly. 

The person who sees themselves in QAnon as a secret hero warrior uncovering the truths is, I think, not that different at all from someone who has that same feeling when playing a game. It’s still the same dopamine, at the end of the day. 

And some ARG players want their games to be real. After 9/11 happened, a few months after The Beast concluded, a vocal minority of our community stepped up and very strongly said they were going to solve the terrorist attacks. My fellow moderators stomped on that impulse: this feeling of achievement, we said, of being powerful and able to solve mysteries was designed so we would feel that way. 9/11 wasn’t designed. Yes, we’d brought together something that felt new and strange — people would call it a collective intelligence — but we’d used it for a maze that had been set down for us by someone who wanted us to solve it. 9/11 wasn’t that. 

This would happen again, and again, and again. In 2013, after the Boston Marathon bombings, a Reddit community declared that it would solve the bombings and find the perpetrator using the power of crowdsourcing. The Reddit community wrongly identified two suspects, one of whom would be later found dead. 

But of course, the issue is complicated and not nearly so simple. It would turn out to be true that the US intelligence community did have a need for more people with Arabic language skills. It would turn out to be true that in some way, a collective intelligence approach might have helped with picking up and amplifying worrying signals. So why not? Why not band together and find that information and do something with it? But it would also turn out, in the 9/11 Commission Report, that those signals were detected and ignored. A collective intelligence, multiplayer game would still, in my opinion, most likely have run into the same systemic, institutional failures. 

I see in QAnon that same ARGish reward of making sense of something and sharing it with other likeminded people. In that way, it’s easy for me to use the lens of game design and see QAnon as a massively multiplayer, distributed, bottom-up, undirected effort that’s strikingly gamelike. Only this game has its tendrils in politics and is a genuine threat to public safety. 

Regular ARGs, the top-down, designed ones with stories, end. When they end well, they end with the players figuring out all of the puzzles and finishing the story. Everything gets wrapped with a bow at least as compelling as the final season of LOST, or at least as definitive. Credits run. When they end badly, they end because the players burned through all the content, all the story, and all the puzzles. 

The problem is, I don’t see how QAnon ends. QAnon is a meme-directed game in the Dawkins sense. It can be understood as a game about an idea that doesn’t really have anyone running it. There’s no singular author, showrunner or writer. There’s not even really a writer’s room. There’s no game designer, no dungeon master. It can make predictions about the world and those predictions can turn out to be consistently, verifiably wrong. What it can do is just keep going and going and going, consuming more links and more information into one giant morass. 

And because it’s not clear that nobody is running it, because the board that it’s played on is the real world, anything that exists in the real world is fair game. Ian Bogost agreed, saying that “in retrospect, the obvious, mainstream endpoint of ARGs was just: the actual internet”. In The Prophecies of Q, The Atlantic’s Adrienne LaFrance realized this too: “if the internet is one big rabbit hole containing infinitely recursive rabbit holes, QAnon has somehow found it way down all of them.” This is echoed by M.R. Sauter in a Real Life magazine piece, too, that “when we impose patterns or relationships on otherwise unrelated things, we call it apophenia. When we create these connections online, we call it the internet.” 

The internet is links, so if there’s something that doesn’t conveniently exist in the real world would otherwise connect your two pieces of red string, you just… make that content exist, and now it’s linkable. QAnon doesn’t run out. It keeps going. A few previous ARGs had tried different strategies to deal with this problem. One, by Jane McGonigal and Ken Eklund was World Without Oil, a serious game where players would collectively imagine and solve puzzles to get toward mitigating climate change. In that game, as in QAnon, fan fiction and user-generated content would become canon, part of the game itself: World Without Oil as an earnest PBS stab at moving us one percent toward a better world, QAnon as a sort of feeding frenzy of pattern-matching.

The origin of QAnon is also, like many ARGs, opaque. Sometimes, ARGs present themselves without context, as pieces of interesting baubles of out-of-place information on the internet designed to catch attention and ensnare people, rather than more explicitly as parts of a marketing campaign as they usually exist now: just like the strange credit for a Sentient Machine Therapist that drew me in in 2000. 

So then a reasonable question about QAnon is whether, in its own language, it’s an “op” — a weaponized, intentionally designed propaganda operation. And through this lens, could it be possible if the people who started it had a history in designing or playing ARGs? 

I don’t think the answer to that question is necessarily helpful or relevant.  Over the last 20 years, the idea of solving complex mysteries has seeped even more into our culture. JJ Abrams, Damon Lindelof and Jeffrey Lieber created 2004’s LOST, which itself had an ARG for viewers to be involved in both during the show’s season and while it was off air. When we launched the Perplex City ARG in 2005, Dan Brown’s massively popular Da Vinci Code had been out for two years. Then Iin 2007, Abrams would deliver his Mystery Box TED Talk, and a more participatory genre of television, mentally involving viewers figuring out a narrative, would become more and more popular. 

(If you do want to go down this rabbit hole, though: some of us in the ARG and game design community did have conversations with DARPA, the U.S. Defense Advanced Research Projects Agency, also famous for instigating the internet. Representatives from DARPA allegedly attended ARGFest, a festival for designers and creators of ARGs in Seattle, in 2013. And in my community’s defense: what if we hadn’t? If the US military wanted to understand ARGs did we want them to understand it from the point of view of optimists and people concerned about the ethical implications, or others who might not care so much?) 

The pool of people who’re familiar with the idea of looking for clues and connections has, I think, only grown over time. The behaviors that we found and had a hunch for, then used to design games in the early 2000s turned out to be prevalent as more people came online and now, to a broad degree, can be relied on. 

The idea that an object could exist, inviting collaborative problem solving and discussion had coalesced in popular culture: it had already leaked out. In many cases, it also collided quite happily with online fandom. 

(I would know: at the studio my brother and I ran, we created an ARg for fans of Muse, an international puzzle-solving treasure hunt, going deep into the works of Zbigniew Brezezinski, as part of the promotion for the band’s 2009 album The Resistance. Fans ate it up). 

But because QAnon isn’t a game— there are just parts of it that look very, very game-like—, it can do things no game can do, or chooses to do, either. 

In the QAnon not-a-game, Bartle’s achiever type, can win something very different than players of videogames like Dota, League of Legends or Fortnite. 

Games don’t let you have a shot at running for congress, for one. On July 3rd, Cameron Peters at Vox wrote an explainer about the QAnon supporters winning congressional primaries; Media Matters reckons that so far there are 9 QAnon supporters running for office in November. While eSports are making money for gamers, if QAnon is a game, its winners are getting ready for public office. 

And our media ecosystem means that achievement translates into social success, too: there are news networks ready and waiting to heap praise and spread their message, from the traditionally problematic Fox News to horrifically unpatriotic upstart One America News Network.

This media ecosystem recognition of players mirrors the design of early ARGs, too. In the early stages of the genre, and following The Beast, it was easiest for ARGs to get funding as marketing campaigns. It followed, then,that any ARG that managed to  include a stunt that would gain media coverage for a player would then be, well, more successful as a campaign, with even more earned media impressions and organic PR. So, then, you could see ARGs as optimizing for more opportunities to create newsworthy events.  

In 2020, QAnon exists in an America where just the idea of a Bush administration official with an unsourced, apocryphal quote about creating reality (“[and] while you are studying that reality—judiciously, as you will—we’ll act again, creating other new realities, which you can study too”) exists. 

The thing is, viewed through the lens of what-if-it-were-a-game, the behavior of QAnon adherents doesn’t look that deranged and crazy. It looks like obsessive fandom for a TV show, or, as others have pointed out, behaviors of certain religions like evangelicism. I think that looking at QAnon from a lens of games, though, might help understand the reasons for the behavior, and why that behavior continues and is rewarded. I don’t think this means QAnon behavior is right, in any way. I do think it makes it understandable. 

My brother and I are hardly the first to have noticed the connection between ARGs and conspiracy theories, nor ARGs and the rise of QAnon. 

Mols Sauter, an Assistant Professor at the University of Maryland College of Information Studies wrote two excellent articles about how online behavior intersects with parts of ARGS in The Illicit Aura of Information in LIMN magazine, and The Apophenia Machine in Real Life Magazine. 

In The Illicit Aura of Information, Sauter analyzed two conspiracy theories and how they treated dumps of data internal emails: the 2009 Climate Research Unit hack (“Climategate”), and the 2016’s hack of Hillary Clinton’s internal presidential campaign (“#pizzagate”, a QAnon precursor). 

In widely-played ARGs, I see reflections of Sauter’s theories of how people treated information in Climategate and #pizzagate. Sauter writes that acquired caches of emails have an illicit aura for three reasons: they gain authority because they’re raw; the act of dumping out information cuts out the role of experts who can confer legitimacy; and they’re relevant because they’re secret.

In 2004, in Perplex City, one of the ARGs I worked on with my brother, I’d write ostensibly private emails between researchers from another world illicitly using a portal to talk about shoes. Players would discover these emails by printing out the teaser website of our game, which would have entirely different content than what was shown on the screen. These clues led players to find a cache of emails that weren’t meant to be discovered. 

Sauter writes that information like this has an aura because it’s raw: and the emails we wrote for our players certainly ticked all the boxes with “imprecise, casual references, professional jargon and elision, in-jokes , and other snippets of not-readily-interpersonal ephemera”. I mean, it’s textbook. We’d even play on the discovery of a cache of emails by having the system they were accessing present them as cached copies of emails that should've been deleted. 

Often, in an attempt to level the field amongst players, we’d attempt to dump out information en masse. Games that had been running for a while would establish gatekeepers, moderators just the role I’d played earlier, to determine what information was in bounds and out-of-bounds. And of course, we’d always play on secrecy as a way to lure players in. 

I used to have a provocative question for my fellow game designers: wouldn’t it be great if there were a rom-com ARG? What would that even be like? Why wasn’t there an Amelie-style ARG? (The closest I feel we ever got was Pemberley Digital’s The Lizzie Bennet Diaries, a sort of transmedia digital show adapted from Pride and Prejudice told through vlogs). 

Now, after reading what others have written and thinking more about QAnon and conspiracy theories on the internet, I’m not sure if ARGs would ever really work outside the conspiracy genre. 

So what are we supposed to do about QAnon? How do we stop it?

QAnon isn’t a game - but when looked at through a lens of alternate reality games, maybe there’s a sort of anti-game we can design to stop it? Or we could use game design techniques to slow the players down, divert them, distract them or render their activity safe. 

I worry that pretty much all tactical interventions won’t work. Because deep down, I think the drivers for QAnon are environmental and systemic. Elizabeth Svoboda at Discover wrote recently about why COVID-19 is turning so many people into conspiracy theorists

Research shows that openness to conspiracy theories comes from causes like low socio-economic status, feeling unsafe due to a lack of agency and control in your environment, and low or negative social connections. 

But what if we looked at QAnon as a piece of media that’s like a game? Then we could see it as an activity that competes for time just like Netflix does, or doomscrolling, or arguing with people on the internet. 

A game design approach to stopping people from “playing” QAnon might be to start making it boring. But again, I worry about the environmental factors. QAnon has to be less boring that the rest of your life. What if the rest of your life isn’t really that great? What if TV and videogames are always going to be interesting than even a more-boring QAnon? 

Put more clearly: you’re in a dead-end job (if you even have one, in our pandemic times). The job prospects in your area aren’t even that great to begin with. You’re socially isolated. Until recently, most things were closed anyway. Government, at all levels, isn’t doing much to help you, and even if it has promised to help you, none of that help has actually arrived. Bills keep coming, because nobody’s helping you out with rent. 

But you could be a winner at this game. 

You could discover that new piece of evidence, that connection no-one else has seen before. 

You could throw it out onto a forum or Twitter or Facebook and get the rush of social approval. 

You get to lose yourself in it because it keeps going, and going, and going. And as you’re doing this and reading and researching, every piece you learn works together to explain the world to you, and explain why the world’s been so shitty to you. There’s actual TV that agrees with it! Everything else? That’s lies. It is as if the story, the hook, the teaser and trailer were evolutionarily selected for disadvantaged and dispossessed people in fear.

I think there are two ways of looking at what to do about QAnon, and they deal with two separate but related issues. The first is that QAnon is merely the latest in a long line of conspiracy theories, and in this way is a symptom of a wider malaise and hurt in our societies. 

The second is the medium of the internet and how it encourages or enables certain innately human behavior. Broadly, I don’t think there’s much the internet can do, directly, about the first issue of the horror of inequitable, indifferent, late-stage capitalism. But there is much that can be directly done to the issue of limiting the metastasizing of conspiracy supporting, harmful group behavior. On July 21, Twitter announced taking “further action” on QAnon, and specifically “behavior that has the potential to lead to offline harm”. This further action included removing 7,000 accounts and limiting 150,000 that had posted QAnon material. 

Twitter could have done this earlier. Reddit could have acted earlier to stop or nip in the bud the community response to the Boston marathon bombings. But those actions come with costs: I feel that in the same way my co-moderators were dedicated to reading every single message that was posted to The Beast’s mailing list (in much the same way that moderatorsproactively manage communities like Metafilter), healthy community management requires consistent and persistent human involvement. There was nobody in the loop, until it was too late. 

But I want to come back to the first point and try not to be too pessimistic about what can be done about QAnon. Because I’ve written before that there are truly good things about the internet: that the same platform of Twitter allows people to truly express what it is to be like them, from discovering and sharing that others share your ability, or inability, to visualize objects in your mind’s eye, through to learning exactly how people perceive color differently. These things are important, at a basic level, because they remind us that we are more similar than we are apart. I’d argue that the internet is still the platform that has the most potential to connect us and build shared understanding. 

What we need, ultimately, is to create safety for people. You could try to fight QAnon with a game designed with benevolent ideas, but those ideas need to be rewarding -- more rewarding, even -- than whatever QAnon can provide. (And remember: QAnon’s rewards aren’t designed in a top-down way. They’re emergent.) 

Arguably, there are not-a-games that do just this, like citizen journalism efforts or, in a very wide sense, the collaborative editing of Wikipedia. But my gut feel is that those experiences are too righteous in their rewards, and not as visceral. 

I think these rewards are going to be incredibly difficult to find and deliver in late-stage capitalism. 

I think the best way to fight QAnon, at its roots, is with a robust social safety net program. This not-a-game is being played out of fear, out of a lack of safety, and it’s meeting peoples’ needs in a collectively, societally destructive way.

In the absence of disruptive, positive social change, I fear the best we can do right now is damage limitation, and of platforms actively managing their communities. 

It’s not impossible. It’s just hard, and we’ve got to want to do it. 

If you made it this far, thanks for reading, and please consider subscribing to support my writing. The customary subscribe button is below.

As ever, I love receiving notes. If there’s enough interest, I may do a follow-up.