On Monday, at WIRED’s 25th anniversary summit, I spoke with Jack Dorsey about some of the biggest questions he confronts running Twitter and Square. How has his position on free speech evolved? Was it inevitable that Twitter would become a tool of international diplomacy? Can the company counter filter bubbles? Should individuals have complete control of their data? And does the internet need a native currency?
Dorsey answered all the questions thoroughly, and sometimes self-critically. Here’s the video of our talk, as well as the transcript:
Nicholas Thompson: I am delighted to get to interview Jack. One of the things about interviewing Jack is that you ask him questions and he does this thing where he listens …
Jack Dorsey: And then I don’t answer.
NT: You usually do answer them! You seem to listen, absorb, and answer. The only thing to do with somebody like that is to ask hard questions.
So I want to start with free speech. Which is one of the issues that would truly divide this room. There are a lot of people who believe that Twitter has abdicated the role that it used to play in defending free speech. You used to be known as the “free-speech wing of the free-speech party,” which I know is a controversial quote, but the idea is that Twitter used to really stand for what may be the most important thing in the US Constitution. And then there are other people who argue that, actually, there are trade-offs between safety, privacy, and free speech. And that Twitter isn’t in the right place on the spectrum. And so I want to just start by asking you for your evolution on this issue, and where you are now on the tradeoffs.
JD: When we started the company, we weren’t thinking about this at all. We were thinking about building something that we wanted to use. And one of the most interesting things about Twitter has been this amazing experiment in creating with others. Everything that we’ve benefitted from—the @ symbol, the hashtag, the retweet, the thread—has been invented by the people using our service, not by us. And it’s been interesting to see what Twitter wants to be and how it manifests itself year after year after year.
Our purpose today, we believe, our superpower, is around conversation. And we believe our purpose is to serve the public conversation. And that does take a stance around freedom of expression and defending freedom of expression as a fundamental human right. Not just one within this country.
‘Our purpose today, we believe, our superpower, is around conversation.’
But it also comes with a realization that freedom of expression may adversely impact other fundamental human rights, such as privacy and physical security. So we believe that we can only serve the public conversation, we can only stand for freedom of expression if people feel safe to express themselves in the first place. We can only do that if they feel that they are not being silenced. So the question we’re starting to ask is: What are the tools we’re building? What is our effect on making it easier to weaponize freedom of expression against someone so that they don’t even feel free to express themselves in the first place or ultimately they’re silenced, which would go against that goal of serving the public conversation?
NT: Is that a different position than you had two years ago? Or do you feel like it is the same position?
JD: Not from two years ago. I think Vijaya [Gadde]’s op-ed in the Washington Post captures our position—our most evolved position—the best. And that is around this balance of people feeling safe to express themselves. But certainly, this quote around “free-speech wing of the free-speech party” was never a mission of the company. It was never a descriptor of the company that we gave ourselves. It was a joke, because of how people found themselves in the spectrum.
NT: But it was a joke that people took seriously and their respect for you increased because of it at the time.
JD: Well, yeah, I don’t think it takes away from our defense of freedom of expression and freedom of speech. But we were not absolute absolutists. A lot of people come to Twitter and they don’t actually see an app or a service, they see what kind of looks like a public square. And they have the same sort of expectations of a public square. And that is what we have to make sure that we get right. And also, make sure that everyone feels safe to participate in that public square.
NT: So the argument you’re making sounds quite a bit like an argument that Kevin Systrom made at one point when I was interviewing him, which is, we don’t optimize for absolute free speech, we optimize for overall free speech and having people feel safe to say what they do. But one thing that they’ve done at Instagram, which you haven’t done, is that they’ve built artificial intelligence systems that can identify bullying, hate speech, cruelty, in the comments. And then last week, they said they’ve now implemented them in the actual images. So they immediately wipe it away without any human intervention. If I say something hateful in an Instagram comment, it can be immediately deleted without me even being told it’s been deleted. I can still see it on my phone, just nobody else can see it. Will you implement similar AI filters at Twitter?
JD: So we will definitely utilize artificial intelligence to make us a lot more efficient and help people feel safe to express themselves. We have been behind in this regard. Our enforcement system mainly relies on reports, which unfairly puts the burden on the victim of any abuse or harassment. So we do need to fix it. So we will be implementing more AI. My worry and what we need to constantly check against is, Can these algorithms that we scale, can they actually explain how they make decisions? And if they can’t, that gets really worrisome.
‘Our enforcement system mainly relies on reports, which unfairly puts the burden on the victim of any abuse or harassment. So we do need to fix it.’
I think this is the most concerning thing to me about AI: this concept of explainability. So there’s a field of research in explainability of algorithms and AI that we have been looking at and intend to continue to research and also help fund where we can. But I do think it is critical, as we offboard a lot more of our decision-making process, certainly from a company but also from an individual level—I’m wearing a watch that helps me make decisions, like to stand every now and then—and what is interesting is that now the algorithms can’t explain why they’re telling me to do that. And I think that is critical for us to, one, trust to continue to give these decisions—individual or company-wide decisions—over and then second, just to understand a bigger picture of what we’re all trying to do at the same time.
NT: I don’t think we solve free speech but that is a good line to draw. Let’s move to another issue that you’ve talked about that interests me. There’s several people this weekend who have argued that one of the fundamental structural mistakes made when the internet was created was around data portability. Our data is mostly owned in central servers by large companies. It doesn’t travel with us. And you’ve actually talked about data portability as something that should be a right. The data should be with the person, not with the company. And you have slightly different philosophies with how you handle it at Square and at Twitter. Explain to me what your ideal version of data portability in the next 25 years would be.
JD: I do believe that individuals should own their data and should have the right to have the controls over how a company might utilize that and how a service might utilize that and be able to pull it immediately. I think a big part of what would make data portability much more scalable is around this concept of federation and all the technologies that something like the blockchain allows for. So I think we are much closer to a world where that becomes more and more possible, to actually have my data rove with me and to be able to be plugged in to any service that I want to use.
NT: But wouldn’t a natural thing that you would do then would be to allow a user to have complete control over their Twitter data. To be able to say, “actually I don’t want you to use it in any of your algorithms; I don’t want you to be able to mine it for information in any way whatsoever.” And I don’t believe that setting exists.
JD: No, we haven’t implemented all of it. But you can see what our understanding of you is from an inference standpoint. You can also turn that off. And you can get rid of individual entries as well. So you can do all that. I think the other thing meant by portability is that I can bring data to it. And we don’t enable that, like we don’t allow an import. We do allow an export and we allow you to turn it off. And in terms of the ranking algorithm, the personalization algorithm, you can now turn that off completely as well. And we intend to go much deeper there, in terms of giving people more control.
NT: I did switch to the chronological feed.
JD: You did switch to it?
NT: I switched and I’ve kept it. I prefer chronological.
JD: Great. You know, ultimately, we want to live in a world where if I turn it off, you’re telling Twitter that you’re not doing your job correctly.
NT: Well the reason I turned it off —
JD: And our job being that, we should help you find the most meaningful and relevant stuff, and if we’re not doing that job, then you should turn it off and you should fire us.
NT: I actually turned it off for, in some ways, even a bigger critique than that. I feel like one of the biggest problems in the world is filter bubbles. And I had a strong sense that the algorithm was accentuating filter bubbles, and that you were seeing people who said the things that you would believe, you were seeing their tweets a lot more. And in the chronological feed, you see people you disagree with. So I turned it off as a way to protest the filter bubbles.
And I’m actually very interested in your views on this because I feel like Twitter makes filter bubbles much worse in the world. But I also know Jack Dorsey grew up in a purple family—a Republican father, Democrat mother—in a purple/red state, Missouri, in a blue city. And I feel like looking at who you follow and the way you act on the internet, you are trying to transcend filter bubbles personally. But I do feel like Twitter makes them worse. So I am quite curious what it is you’re doing to counter the breaking of American society into filter bubbles?
JD: I think Twitter does contribute to filter bubbles. And I think that’s wrong of us. I think we need to fix it. But I don’t think it’s the chronological timeline or the ranked timeline that does it. I think it’s the fact that we only enable you to follow an account.
NT: So how do you fix the filter bubbles? I will switch back to algorithmic timeline the minute you implement these next things that you’re about to tell us about.
JD: So if I’m following an account of someone who may have a particular viewpoint, that account may be only talking about their particular viewpoint because of their particular bias. [Contrast that to] the ability to follow a topic, or an interest, or an event—like, during Brexit, if I was to follow @vote_leave I would likely see the majority of it being Vote Leave, but there’s probably a few tweets that counter it. And right now we’re not even showing those because most people are following the @vote_leave account and the Vote Leave people. And we’re not giving them the tools to even have an opportunity to break down the filter bubble. But that has nothing to do with the chronological timeline.
‘I think Twitter does contribute to filter bubbles. And I think that’s wrong of us. I think we need to fix it.’
So, we’re looking at what our fundamentals are that we started 12 years ago that made sense then that may not make as much sense today. And that’s everything from the tools we give people, like following an account, which is the only thing you can do, to when you open up the app, what is it incentivizing you to do? And should it incentivize you to do anything? And this is where, you know, the likes and the follower count and retweet and reply come in.
NT: I love this answer and it gets at one of the things I’ve wondered the most which is the unintended consequences of tech. You created this tool for communicating with your friends and for letting your mother know what you ate—which I’ve listened to some podcasts and you talk about wonderfully—and it’s become a tool in which the president of the United States threatens nuclear war. The delta is large.
JD: It’s become one of the mediums that he has used to do that.
NT: The delta between Twitter as described by Jack Dorsey in podcasts 10 years ago and Twitter as used today is large. Was that inevitable, that a short-form messaging service used by millions of Americans would become this? Or was it the result of the code? I’m curious what you think about that.
JD: That’s a good question. I don’t think it was a result of the code per say, I think once people saw the possibility of—I think Marc Andreessen said this amazing thing that I still believe today, which is this concept of global instant public messaging for free. I mean, if you had this app in front of you that was free that you could actually text message the entire world, that’s what Twitter is. And once the world saw that, there was no taking it back.
So whether it was Twitter doing that or some other service that allowed for a lot of very public, open conversation without this “join” button or “leave” button, anyone could see, anyone could follow—once the world saw that, it is what it is. And it’s always going to be there. So I think the world, again, once they saw it, they needed it. Our job now is to make sure that we’re actually serving that and not just assuming that we’re it.
You know, there is a possibility for a more global public conversation. I think there’s a need for it. You know, someone you’ve talked with in the past, Yuval [Noah Harari], said in his most recent book, 21 Lessons for the 21st Century, the problems the world is facing can only be solved by a global conversation and global entities. The organizing principles of nation-states cannot solve things like climate change and it can’t solve things like economic inequality, and it can’t solve things like nuclear war or displacement of work by AI. So we’re either going to react into a global conversation, or we’re going to start having more of them proactively. So we need to make sure, as a company, that we’re providing a service that at least makes that potential stronger.
NT: So the stakes for Twitter are even higher than the way I laid them out before. One of the other points that Harari makes is that the platforms incentivize certain conversations and certain behaviors, and so if you’re going to create a global conversation about how to deal with displacement of AI, it’s pretty important to get the fundamentals right. So I’m glad we’re talking about it today. And I’m glad you’re thinking hard about it.
‘We should always be questioning Twitter.’
JD: Absolutely. And also just question what the app and the service currently incentivizes. Because right now we have a big “like” button with a heart on it, and we’re incentivizing people to want to drive that up. We have a follower count that was bolded because it felt good 12 years ago, but that’s what people see us saying, that should go up. And is that the right thing? Versus contribution back to the conversation or contribution to the public conversation or healthy contribution. How do we incentivize healthy behaviors? Realizing that not everyone’s going to choose health in the moment. For the collective, how do we help people choose things that might contribute more to a global conversation?
NT: it sounds almost like you’re a journalist questioning Twitter there.
JD: We should always be questioning Twitter.
NT: We only have a couple minutes left and I want to talk a little bit about money and finances, since, extraordinarily, you run two companies. One of the things that you’ve said in past interviews that I’ve been struck by is that there should be a universal currency on the internet. That the internet deserves some kind of payment system, and you think it could be bitcoin. Where do you stand on this issue right now?
JD: Well, I fundamentally believe—and you know, I grew up on Usenet. I grew up in St. Louis, Missouri and I was bored out of mind. I went to Usenet, I found WIRED when the William Gibson cover came out, so thank you very much. I’ve always been a fan, but I lived through Usenet, and one of the most interesting Usenet groups was alt.cypherpunks, which WIRED covered a lot back in the day. It was the most edgy, interesting, kind of cult within the internet at the time, and I think still is. And the dream was this native currency for the internet. And we have so much that was born on the internet, and grew up on the internet, but we don’t have that currency. And you look at what the internet has done to, again, Yuval’s concept of these organizing principles of a nation-state, it has gone above that. And it has made that meaningless. And one of the last remaining is currency.
If we are to have an internet population—if we are to have internet-based media—we will need an internet-based currency, and a native currency at that. And I don’t know if we’ll have many native currencies or one will be chosen, but I do know and feel strongly that the internet will have a native currency, if not a collection of them. And I just want to take on the mind-set of, I don’t want to wait for the thing to happen, I want to help make it happen. Because I think it does continue to get us all together facing the same existential problems that we all need to realize we need to face as a global population.
NT: It would also do a public service for Jack Dorsey. If we had a global currency that makes the global conversation work better together, then your two companies would make more sense and all 16 hours of your day working would be spent working on the same project. We are out of time, thank you very much Jack Dorsey for joining us. It’s been wonderful to chat with you.