Susan Benesch is the Faculty Associate of the Berkman Center for Internet and Society at Harvard Law School and Director of the Dangerous Speech Project, where she has a built a framework to identify dangerous speech and diminish its violent effects, while upholding freedom of expression. In India for a public lecture and to begin work on a three-year study, she spoke to Rukmini S. about the Internet as a means to disseminate dangerous speech content. Excerpts:
You talk of dangerous speech as a subset of hate speech, with a particular focus on whether it is able to incite violence. What is your thinking on the harm from speech that might not be a call to violence but still affects a minority or other group?
Dangerous speech is a subset of hate speech because there is hate speech that doesn’t seem to increase the risk of violence. There is also speech that is dangerous but not necessarily hateful; it is possible to increase the risk of violence by telling people something that isn’t at least explicitly hateful. For example, if I say another group of people is planning to attack your group, particularly in certain circumstances that could be highly dangerous, especially if it seems believable. But it doesn’t necessarily include hate.
Another reason why I’m careful of and even suspicious of the term hate speech is that I’m not sure that hate is always the most important emotion for turning one group of people against another; fear is also very important.
I define dangerous speech in order to do several things. One is to protect freedom of expression. Hate speech is a variously defined term; there’s no consensus definition. In the breach, it tends to be defined very broadly, and laws against hate speech are very often used by governments to crack down on their own people — a lot of journalists in many countries, for one.
The second thing is to get people to focus on some of the worst consequences of some speech and especially on the consequences against which we have some pretty good consensus across groups and countries. Violence — especially large-scale, inter-group violence — is one consequence at one end of the spectrum. There are many other kinds of harm, though, as your question implies, that speech can foster (I don’t necessarily want to say “cause” because it’s hard to draw a causal line between speech and action).
Also, the question of whether speech is dangerous is not a toggle. Speech is not either zero dangerous or 100 per cent dangerous — speech can be more or less dangerous. So it’s a spectrum already within dangerous speech. And then if you look at hate speech, there’s a broader spectrum. So plenty of speech can, at a minimum, make people who it purports to describe feel diminished, humiliated, frightened. As the speech makes them feel more and more frightened, usually it is because they are perceiving a more and more real threat since at the same time it makes other people feel even more willing to condone or participate in violence.
Is there something in particular about the Internet that you feel lends itself better to being a means of dissemination of dangerous speech? Can the fact that messages are disseminated online make them more dangerous?
Yes, I think so in various ways. One, people can be in contact with far more people and people who are further away. ISIS is extremely effective in using online means of dissemination and they are able to gather people who, aside from the connections they are making online, would otherwise feel themselves to be very diverse geographically, if nothing else.
The fact of feeling more connected as a group can make a message more effective. The second way in which the Internet can perhaps not increase the dangerousness of speech but allow for easier dissemination of more dangerous speech is anonymity, and particularly what they call the disinhibition effect. So a young person who receives a dangerous message is more likely to forward it online than he or she would be to repeat it to his or her mom. That shifts the discourse norms; it becomes ok to say certain things online that you would know not to when you’re offline.
If you do think that the Internet lends itself to more dangerous speech, do you think it then calls for more robust policing?
I’m quite sceptical of law enforcement for a few reasons. One is that the traditional methods that all human governments have used to suppress speech that they didn’t like have been only two — punishment of the speaker and censorship of the speech. That’s it: two small tools. So, online censorship means deleting and the only even remotely effective way of doing that would be prior censorship. For example, Microsoft has a system like that for Xbox — there’s a huge database of words and phrases that you can’t put on Xbox; they’ll disappear immediately. I’d be terrified by that prospect if Facebook does it, if a government does it, if anybody does it, because we don’t get to see the list.
Post facto censorship is what we have then, and that’s Facebook trying to delete stuff. Some of that is of course necessary — with child pornography, for example, or images of violent rape. There are many examples of online content on which we would agree that it’s correct and proper for someone, whether it’s an Internet platform or a government or even both, to do the best they can to take it down as quickly as possible. However, hate speech and dangerous speech describe a large spectrum and there are so many examples of speech that some people and some communities find very offensive and dangerous and others won’t.
The other thing about post facto censorship is that first, it doesn’t necessarily correct the harm. The big Internet companies are not policing the content independently except for some exceptions like child pornography and copyright violations. They are waiting for users to report content and then they look at it. When any Internet platform or law enforcement learns of something that has been posted online, by definition it means that someone who is at least offended by it and may be harmed by it has already seen it. So, some harm has already taken place. Now they’ve deleted it, but there’s no guarantee that the person won’t post the same thing again… Some of the people who are producing this stuff are only invigorated by attempts to delete. You can prosecute people, but I’m not convinced that the prosecution will diminish their speech or its harm…Sometimes prosecutions can only make their targets more famous.
Rather than going the law enforcement route, one of the strategies you propose to take on dangerous speech is “counterspeech”. How do you picture this playing out — armies of volunteers who jump into Twitter conversations to counter dangerous speech, for example?
I’m doing research on that right now on a two-year project called the Kanishka Project, with a computer scientist at McGill University. We are hunting for those admittedly rarer cases, but nonetheless existing ones, in which this is already happening — spontaneously or naturally — online. We’d like to learn about how it is that dangerous or hateful speech is successfully countered online.
Secondly, I’m happy to say that other people have already started some wonderful projects that seem to be working quite well. For example, the Panzagar project in Myanmar. For the last several years in Myanmar, there has been an increasingly vicious and frightening campaign of propaganda against Muslims, being led by various people including, I’m sorry to say, some influential Buddhist monks. These messages are disseminated in various different ways including in offline speeches, and Facebook is also a very important means of dissemination. Some mainly Buddhist activists in Myanmar decided to do something against this and push an opposite message online and offline.
They’ve had a campaign going now for a year and a half in which they travel around the countryside and hold meetings, and pass out stickers and notebooks for children with their peaceful message that is embodied in the name of their campaign, which means Flower Speech or Flower Language. Their page is not as rampantly popular as the extremist pages but they’ve had quite a lot of success and I, for one, am happily surprised.
I don’t pretend that counterspeech can bring around a lot of people who’ve spent a lot of time disseminating hateful, violent, fear-engendering messages online. What I do think is possible is to shift the discourse norms, to influence what all the other people, the audience online, thinks is acceptable. That may sound idealistic, but it isn’t. We do have substantial evidence that that’s happening successfully. I ask my students in the United States some times: what is the likelihood that an American political candidate would use the n-word? It would just not happen, and that is even though the use of that hateful word is protected by our First Amendment.
Discourse norms can shift within any human community frighteningly quickly when they’re going the wrong direction. It has become much more acceptable to say awful things about Muslims in many countries, including mine. They also can shift in the right direction astonishingly quickly, and that’s another important thing to remember about online communication.
Is there anything that you’re working on to do with dangerous speech in India?
I’m very excited to say that the Berkman Center for the Internet and Society that I’m attached to and the National Law University Delhi’s Centre for Communication Governance are beginning a collaboration to work on these questions of hate speech and dangerous speech and especially what can effectively be done about them while also protecting freedom of expression. India is a particularly important and exciting country to do this.
There’s obviously a problem, as there is in so many other places. There’s also a tremendous critical mass of people working in brilliant and important ways against it. It’s a three-year study and nothing on that scale has been done on online hate speech anywhere in the world.
