On the (In)security of Social Media Callouts

If your first instinct upon reading the title of this blog post was the Ctrl+F for the words “SJW”, “woke”, or “cancel culture”, you’re already assuming incorrectly about this post and the author’s positions.

If “please don’t do that” is asking too much, maybe sit this one out?

We like to think we live in a civilized society. This implies that misbehavior is appropriately addressed by professionals whose goal is to reduce harm and promote liberty. A justice system, if you will, that prioritizes guilty people walking free over innocent people suffering. One that emphasizes restorative justice over imprisonment and ruining people’s lives forever.

Unfortunately, that’s a pipe dream, and the actual justice system is so deeply flawed that basically half of John Oliver’s Last Week Tonight skits are about how fucked everything is.

So right off the bat, any discussion of social media callouts that hinge on the criminal justice system in a given country are probably not worth having. Garbage in, garbage out.

Presumed innocent until proven guilty, right to due process, etc. are legal concepts (which often don’t even exist in practice; ask your BIPOC friends about it sometime) that fools often invoke in order to shield abusers from public accountability.

Whether or not a person can be effectively prosecuted for their misbehavior depends on a lot of factors (most of which justice should be blind to, but isn’t). However, social consequences are not bound by a courtroom. Public opinion and social accountability are orthogonal to Law.

So to be very clear, what I’m talking about today are callout posts on social media (i.e. Twitter), and the social consequences thereof. I am not talking about the legal system or the responsibilities of a free press.

Understanding Social Media Callouts

To explain by analogy, social media callouts are the digital equivalent of the disapproving glare from meatspace. (Of course, it’s not a perfect 1-to-1 mapping.)

The framework for a social media callout is predicated by this social scenario:

1. A person or company (X) says or does something harmful (H) to a community (Y).
2. X is escorted out and disinvited from future events by Y until they correct their behavior and make amends with the people they hurt.
3. If and only if X has corrected their behavior and made amends with their victims, they can be reintegrated into Y.

Not only is this reasonable, but it’s necessary. Tolerating misbehvaior (negative peace) or intolerance leads to social toxicity, which hurts the entire community.

However, social media makes things much more complicated, because communication is distributed and asynchronous.

In a stark departure from the predicate framework, a social media callout looks more like this:

1. A person or company (X) says or does something harmful (H) to a community (Y).
2. One or more witnesses (W₀, W₁, …, W_n) observe the harm in step 1. (Some or all witnesses may be victims of H, but don’t have to be. They’re also not guaranteed to be a member of Y.)
3. Some or all witnesses decide to publicly decry X for H on a social media platform. This is the callout.
4. Members of Y are made aware of H, and signal-boost the output of step 3 to other members of Y.
   1. Since there is no effective distributed consensus mechanism (note: please do not fucking recommend a blockchain), awareness of H isn’t uniformly distributed through Y.
   2. Many members of Y will choose to simply block X in response to being informed of H.
5. If and only if X corrects their behavior and makes amends with their victims (~H), they can petition the diffuse collective of Y to forgive them.
   1. There’s no reliable mechanism to verify that the preconditions even satisfied, at scale.
   2. There’s no reliable mechanism to ensure all members of Y are aware of ~H, at scale. This is especially true if they blocked X.

Even worse, this is ignoring the influence of the social media engagement algorithms that will expose a different portion to H than ~H, even if nobody chose to block X for H.

This is also an oversimplified model of the real world. There are usually even more steps, and a lot of ego involved, which leads to one or more parties doubling down on their points (whether or not they’re valid or proportionate).

This is already a messy situation, with lots of room for error, and I haven’t even gotten to the point I’m trying to make.

Malicious and Inauthentic Behavior

Internet trolling is a term used to describe a deceptively broad range of behaviors. You have shitposting, mostly-harmless pranks, malicious compliance, and myriad memes, on the one paw. But it also describes obsessive stalking, hateful or violent conduct, coordinated inauthentic behavior, deliberate misinformation, gaslighting, astroturfing, sockpuppeting, doxing, and generally being a fucking nuisance.

In the overwhelming majority of social media callouts, the actual risk a community faces is that the original offender (X) and their supporters will engage in trolling behaviors. Their incentive here is pretty clear: Regardless of the validity of claims about the harm done (H), they want to discredit the callout. This often means bullying or discrediting witnesses (W_i) and/or victims.

However, it’s not the only threat that communities face with social media callouts.

This bears emphasizing: There is a very stark difference between a risk and a threat. A threat is something hypothetical and bad. A risk is a statement about the probability of a threat happening, and its potential consequences.

With that in mind, here are some threats that a troll can leverage to cause harm through a callout (which are obstenisbly intended to reduce harm).

An Incomplete List of Threats

1. False Callout.
   The simplest attack is to make a dishonest callout, falsely accusing your victim of some imagined harm. This usually looks like DARVO when it happens.
   
   This is pernicious because you can split the room between the “believe victims” camp and the “check receipts” camp, and cause the community to devolve to pointless infighting and philosophical wankery instead of helping anyone.
   
   (We also don’t know how frequent this threat actually gets exploited in practice, but heuristically it’s less common than a cover-up for a true call-out.)
   
   This can also be done in retaliation of another callout, or as a mob harassment strategy. This can ruin lives and make people feel hopeless and isolated, which can increase the risk of suicide.
   
2. Partition Sybil Attack.
   The inequity of the social graph allows the existence of perceived local majorities, wherein community members find themselves surrounded by people who believe in the innocence or guilt of X, and will loudly pressure their peers to act accordingly.
   
   This can lead to bandwagons (agreement) or ostracizing (disagreement).
   
   In practice, this is often manifested as gaslighting or astroturfing, and may involve sockpuppet accounts.
   
3. Alert Fatigue.
   Emotional bandwidth is a scarce resource. Resource exhaustion techniques can lead to alert fatigue, which results in poor judgment.
   
   By overloading a community with callouts of dubious quality, people will either grow apathetic or hypervigilant, and then identify very strongly with their stress response and infight with people that have different stress responses.

There are probably many others that I’ve overlooked. It’s difficult to assign probabilities to these threats and make an informed risk assessment, but they’re obvious consequences of even the simplest social media callouts.

Okay, So What’s Your Point?

Most treatments of social media callouts fixate entirely on the validity of the accusation (threat 1), and often conflate community responses to vastly different types of callouts in an incredibly dishonest way. However, I think there are other threats that online communities have largely ignored.

If your response to someone sharing a meme that originated with a probable appropriation of AAVE is the same as your response to sexual abuse within your community, you’re part of the problem.

Be intolerant of Nazis? Absolutely! The paradox of tolerance requires it. But it doesn’t specify how you should practice this necessary intolerance. On this topic, Karl Popper’s observation is inert.

There probably isn’t a one-size-fits-all solution for this complex problem space. Social media algorithms may make this worse, but humanity is poor at distributed consensus even without them.

It doesn’t help when two camps are polarized, and trying to find a happy compromise is basically impossible.

<-------------- Response Magnitude -------------->

 <=========={ Not Far Enough }==========)
         (=============={ Too Far }=============>

[__________________________^_____________________]

Of course, I do have opinions on what should be done differently, but so do many (if not all) of you who read this post.

And as tempting as it may be to make this post about me, or my experiences with false accusations of shitty behavior over my lifetime, or my opinions about how social justice should be done “correctly”, I don’t actually have answers to any of these problems. I don’t even know that the problems are even solvable in the first place.

We live in a world of unmitigated threats and imperfect communication. The biggest act of hubris is probably thinking we’ve got it all figured out, and that it’s obvious, and anyone that doesn’t agree has a secret agenda to enable abusers and/or false callouts.

We all make mistakes (especially with wording), but most of us don’t commit heinous acts (sexual assault, violence, hate speech, donating money to bigots, etc.).

Maybe we should be more forgiving for the former, and less tolerant of the latter (if, for no other reason, than for the safety of our communities).

Maybe, for non-heinous acts, we could try to settle disagreements privately (with a trusted mediator, if emotions are running high) instead of airing our grievances all the time.

Maybe we could make more of an effort to engage in good faith, ourselves, even if the other party is not.

For full transparency, I block a lot of people, including many furries, on Twitter. That doesn’t always mean I think they’re beyond redemption. I block them to assert my personal boundaries, and to preserve my scarce emotional bandwidth for the people I’m close to. I don’t have the time or energy to, for example, help deprogram former “alt-furries” from the pewdiepipeline.

To Answer the Meta-Question

No, nothing recent happened on Twitter to prompt me to write this. Thankfully, despite receiving international media attention for leading the effort to save a library from a homophobic mayor, I haven’t been drowning in conflict. As far as I’m aware, nobody’s trying to “cancel” me today.

This is just a topic that’s been floating around in the back of my mind for years, and it gets even muddier every time someone misbehaves.