Re: Against Stochastic Terrorism
What is stochastic terrorism?
As per Scott’s article:
Some other popular examples of the concept:
“Stochastic terrorism” is the idea that if you spread fear and mistrust against a target, then eventually people will commit violence against that target, and it will be your fault, even if you never specifically said the words “you should commit violence”.
Nativists spread fear and mistrust about Muslim immigrants, and then racists go on shooting sprees in mosques.
The #Resistance insists that Donald Trump would destroy democracy, and then various people try to assassinate Donald Trump.
Conservatives spread fear and mistrust about transgender people, and then bigots commit hate crimes against transgender people.
Woke people say the police are racist and brutal, and then other woke people murder police officers.
Socialists call health insurance companies greedy and accuse them of blocking life-saving treatment, and then Luigi Mangione murdered a health insurance CEO.
AI safety activists say that Sam Altman’s AI could destroy humanity, and then a guy threw a Molotov cocktail at Sam Altman’s house.
Liberals said that Charlie Kirk was a hatemonger who was driving Americans apart, and then an assassin murdered Charlie Kirk.
The pro-life movement describes abortion as murder, and then pro-life activists assassinated abortion doctors and pro-choice politicians.
A weak post, below Scott’s usual standards.
He begins by tainting the concept by association:
“Stochastic terrorism” mostly gets deployed opportunistically, by people who either are too blinkered to realize that the same argument could be leveraged against their own speech, or who hope you’re too blinkered to realize that.
True, but irrelevant to whether it’s a useful concept.
He then argues no consistent principle can separate normal criticism from stochastic terrorism. He proposes weak candidate principles, knocks each down, and overreaches to the “liberal solution”:
The liberal solution is none of these: it’s that it’s always legitimate to criticize any person or group however strongly, and never legitimate to use extra-state violence against them; if a violent act occurs, it’s 100% the fault of the perpetrator, and 0% the fault of other people who previously criticized the victim.
Convince readers of a state of the world—true or false—and you bear responsibility for the reasonable reactions it invites.
If you convince a man he can fly and he throws himself off a cliff, you are responsible for his death.
Say “Sam Altman’s AI will destroy humanity” and assassination becomes rational to anyone who believes it. A man about to end the world is a movie villain holding a doomsday device. Assassination, kidnapping, and airstrikes all beat the alternative.
To a great many people, the reasonable, rational thing to do would be also the morally correct thing to do.
If readers then believe X will escape justice or continue unchecked unless they act, “morally correct” becomes “morally necessary.” That gives you the ingredients needed to turn stochastic terrorism into real violence.
(If you claim Trump will enact concentration camps, you imply Congress and the courts won’t stop him, while the military, federal agencies, police, and contractors play ball. What option remains but violence?)
Don’t shed responsibility for your claims. Calibrate them.
- “Trump is destroying democracy.” → “Trump is weakening democratic norms.”
- “Sam Altman’s AI will destroy humanity.” → “Sam Altman is increasing AI existential risk.”
Responsibility attaches to the claim you made, not a distortion of it. Whoever distorts it owns the distortion.
Incidentally, this same logic makes Trump responsible for the January 6 Capitol riot. Though he did not directly order it, his claims made rioting seem like a reasonable response.