[ad_1]
For the first time in human history, we can measure this thing a lot with tremendous accuracy. All of this data exists and is constantly being evaluated by companies. What are the consequences of our rules? Every time they make a rule they test its enforcement effects and possibilities. The problem, of course, is that everything is blocked. No one has access except the people of Silicon Valley. So it’s very exciting but also very frustrating.
This connects perhaps the most interesting thing for me in your paper, which is the concept of probabilistic thinking. Much of the coverage and discussion of content moderation is based on anecdotes, just as humans do as usual. Like, “This part of the content, Facebook said it wasn’t supported, but was viewed 20,000 times.” One point you mention on paper is that perfect moderation of content is impossible on a scale unless everyone is banned from what no one wants. You must accept that there will be an error rate. And each option is the error rate in which direction you want to go: do you want more false positives or more false negatives?
The problem is if Facebook comes out, “I know it looks bad, but the truth is, we’ve removed 90 percent of the bad stuff,” which doesn’t satisfy anyone, and I think that’s one reason we’re stuck taking the words of these companies.
Totally. We have no idea. We stopped at the mercy of that statement in a blog post.
But there is a grain of truth. Mark Zuckerberg has this line that is spreading all the time in all the testimonies and conversations in Congress. The police don’t solve all crimes, you can’t be a crime-free city, you can’t expect perfect enforcement. And there is the issue of truth. Moderating the content to be able to establish order to the confusion of human expression is a dream idea and it is quite frustrating, unreal and fruitful about the constant stories we read in the press. Here is an example. an error or set of errors, that this rule cannot be properly complied with.
The only way to achieve perfect enforcement of the rules would be to ban anything that looks like this from a distance. And then we would remove the onions because they look like boobs or whatever. Maybe some are not so concerned about free expressions for onions, but there are other worse examples.
No, like someone who watches a lot of cooking videos—
It would cost a lot to pay, right?
I see a lot more pictures of onions on the net than breasts, which would really hit me.
Yes, exactly, so it is a strong cause of free expression of onions.
I’m into that.
We have to accept mistakes one way or another. So the example I use in my paper is in the context of the pandemic. I think it’s very useful because it makes it very clear. At the beginning of the pandemic, the platforms had to send their workers home like everyone else, which means increase confidence in the machine. They didn’t have so many humans doing the checking. And for the first time, there were really clouds around the effects of that, which is, “Hey, we’re going to make more mistakes.” They usually come out and say, “Our machines, they’re so great, they’re magical, they’ll clean up all those things.” And then they said for the first time, “By the way, we’re going to make more mistakes in the context of the pandemic.” But the pandemic opened the way for them to say that, because they were all like, “Okay, make mistakes! We need to get rid of this thing. ”That’s why they were wrong to get the wrong information because the social cost of not using the machines at all was very high and they couldn’t trust humans.
In this context, we accepted the error rate. We read stories in the press about how masks were bad at the time, and they banned mask ads, their machines necessarily added that, and they also threw out a lot of voluntary mask manufacturers because they were like machines. , “Bad masks; drop them. ‘ And okay, it’s not ideal, but at the same time, what choice do you want to make there? On the scale, while there are literally billions of decisions, all the time, there are some costs, and we were freaking out about the mask ads, so I think it’s more reasonable to do that.
[ad_2]
Source link
