From the August 8 edition of Premiere Radio Networks' The Sean Hannity Show:
JACK DORSEY: Thank you Sean, first for the opportunity to talk to your listeners and also painting the picture of the complexities that we’re facing. This is definitely not easy but we’re trying to approach this with very simple principles; how do we earn more trust?
These are models that are looking at behaviors and behaviors of bad faith actors who intend to manipulate, distract, divide a conversation or to unfairly amplify their content which they didn’t earn. So those are the signals that factor in and we do rank search, we do rank trends, and we do rank conversations accordingly. That does not effect one’s timeline, if you follow someone on Twitter, you’re going to see them in your timeline.
SEAN HANNITY (HOST): What about, and I think this might be something everybody agrees on, let’s start with people that are calling for some type of violence of any kind or threatening violence against an individual, I think that would probably be an easy, we’re not going to allow that, right?
DORSEY: Yes, that is much easier. Any sort of violent speech, encouragement towards violence, harassment, is directly against our terms of service and we take immediate action on it.
HANNITY: What do you do in the case of a figure that is known either a white racist or supremacist or a KKK member or maybe on the other end of the spectrum you’ve got Louis Farrakhan, known anti-Semite, racist, what do you do when they set up an account? Is it that you monitor them and if it’s a certain video that they’re linking to, if it’s a certain article that has racist ramblings in it, then it becomes a little, again, it becomes a little more nuanced sometimes. Or maybe somebody is an overt racist and they use horrible language and then maybe somebody is more subtle and they use code language that is racist. In that sense it becomes harder, does it not?
DORSEY: It does become harder but so we rely on a bunch of signals including reports from those who that account might attack or from bystanders and then we also take into consideration again, some of the context of everything that’s happening around it. Some of the groups that you mentioned earlier on, there might be a violent extremist groups that try to get onto to our service and we take that into consideration. We also look in those particular cases at off platform behavior as well. So things that aren’t just happening on Twitter, but happening on other platforms or in the real world.