Twitter's Former Head of Trust and Safety Finally Breaks Her Silence
wired.com
wired.com
"This time she’s sending a screenshot of tense tweets exchanged between the supreme leader of a nation-state and the official Twitter account of a rival nation-state. I respond with a prompt: “Let’s say you’re still at Twitter. How would you handle?” Harvey walks me through her decisionmaking process, then turns the tables: “What would YOU do?” I ask her whether I can sleep on it. She tells me I do not get the luxury of sleeping on it. No matter which call I make, people will despise me for it. And no matter what the outcome is, people could die. “You have to live with making the least bad decision you can with the resources you have,” she writes."
I don't understand, someone is arbitrarily deciding who can and can't write things on a platform made for people to write things on? Maybe I don't have the capability of thinking deeply about this kind of situation yet, but it seems pointless to me. Whether it's Twitter, newspaper headlines, a televised speech during a press conference, the same thing can happen. The whole article seems like useless drivel about a useless role irrelevant to the interests of this site. Isn't hacker news all about sharing the latest and cool new breakthroughs in technology like batteries, the new Apple Silicon hardware, the new rebar technology available starting from the nVidia 3XXX-series?
These are just tweets. Like a few short sentences. The world is a big place. I'm sure someone has died once because of some event connected to a tweet, but it can't happen very often. Right? Or am I just naive?
For each of these cases that do then occur, approximately 10% of them require somebody representing the platform to attend court (usually briefly) and approximately 1% of those result in a judgement against the platform even when they show that the platform took all reasonable precautions and legal requirements including, where applicable, closing accounts and reporting inappropriate behavior to the relevant authority.
So let's run the numbers for one million US users. You have 50 real psychos on the platform. Roughly 5 violent assaults (or worse) connected to the platform occur a year. 0.5 of them require the platform to send a legal representative.
Assuming the platform is following all legal requirements and can show this, there are 0.005 legal judgements in the USA against the platform for enabling violent crime per year for every million US users.
Or, if you have 200 million US users, you can expect 1 court judgement for not doing enough about violent criminal activity a year. Now, what about the four other major categories of crime communicated by US residents on the internet: drug sales, child sexual exploitation, domestic terrorism and piracy of entertainment products? That 1 judgement a year, unless you have a very good legal team, will become 30 judgements a year.
When they say that "people could die" they are using an attention-grabbing shorthand for "if we can not demonstrate a commitment to full legal compliance then the headquarters will be raided by two dozen FBI vans, or, if we're unlucky, Secret Service vans." Note that appearances matter much more than results, particularly because of the senility of the average policy maker in the United States. Look how badly Meta has had to drop the ball to get 9-digit fines and they have ten times the userbase of our hypothetical.