Can Candid Conversations Happen Online Without The Trolls? | KUOW News and Information

Can Candid Conversations Happen Online Without The Trolls?

Aug 1, 2016
Originally published on August 1, 2016 3:37 pm

A few years ago, Silicon Valley engineer Bindu Reddy was raising money for a new startup. An investor offered to contribute — not because of what she was trying to do, but because she was a woman.

That rubbed Reddy the wrong way, and she wrote about it — then the backlash began.

Now Reddy's goal, with the new social network Candid, is to facilitate online conversations but without the trolls. She spoke to NPR's Kelly McEvers about finding the balance between free speech and moderation on social media.


Interview Highlights

On the dangers of online speech

As a person who's in a professional field, you have to censor your opinions. And if you don't, you get attacked by trolls, and sometimes you get judged negatively by your peers, and that might even affect your career. And what I've seen again and again is a lot of my friends and my peers kind of hold back and not speak openly.

On distinguishing Candid from anonymous platforms like Yik Yak and Secret

Now, all of these other platforms, they were founded a couple of years ago. Over the last two years, there's been a lot of advances in natural language processing [NLP], in machine learning, and kind of artificial intelligence that the machine can understand what you're saying. ...

The algorithm on Candid is a learning algorithm, so as we get more data we learn more. But the idea is to kind of weed out the bad posts, as I call them.

On how automated moderation works

The way other artificial intelligence works, is it basically parses your sentence. We use a deep learning NLP algorithm, which basically looks at what you're saying and decides ... whether it's positive or negative. So it kind of classifies things as having a negative sentiment or a positive sentiment.

It then gives it a score of how kind of strong your statement is — let's say you said something about someone or you threatened someone, it classifies that as saying, "Hey this is a very strong statement," because these kinds of categories are not good in terms of social discourse. And when we do that, we basically say if this thing has a score which is more than a particular level, a cut-off, then we basically take out the whole post. So whether it's self harm or like bullying or harassment, we look for certain phrases and the context of those phrases.

On the line between moderation and censorship

I mean, here is the thing between what is "loud free speech," quote-unquote, right? At some level you should be able to say what you want to say, but on another level, you also want to facilitate, you know, what I would say constructive social discussion. ...

There is a kind of a trade-off or a fine line that you need to walk, because if you let everything in, you know the fear is that social discussion stops and it just becomes a name-calling game. And that's what happens if you just leave — like certain discussions, just let them be, don't pull things down — you will see they quickly devolve into people calling each other names and not having any kind of constructive conversations.

Copyright 2016 NPR. To see more, visit http://www.npr.org/.

KELLY MCEVERS, HOST:

On this week's All Tech Considered, the story of a new anonymous social network called Candid that claims to be a space where hate is not welcome.

(SOUNDBITE OF MUSIC)

MCEVERS: Sometimes you say things you wish you could take back later. If you say them on the internet, they have a way of living forever. That's what happened to the founder of Candid, Bindu Reddy. A few years ago, she was raising money for another tech startup. An investor offered to contribute simply because she was a woman. That rubbed Reddy the wrong way, and she wrote about it, then came the backlash. Reddy told me it made her regret posting her thoughts online.

BINDU REDDY: As a person who's in a professional field, you have to censor your opinions, and if you don't, you know, you get attacked by trolls and sometimes you get judged negatively by your peers. And that might even affect your career. And what I've seen again and again is a lot of my friends and my peers kind of hold back and not speak openly.

And Candid is really, you know - has been created to address that issue. It's, you know - it's a way to kind of express yourself and express your thoughts and opinions without having the negative consequences of, you know, speaking out in social media without, you know - without facing hate and without being attacked.

MCEVERS: You know, other sites have tried anonymity before. You have Secret, Whisper, Yik Yak, you know? Users have accused Yik Yak of allowing for harassment, rumors, bullying, threats of violence. I mean, this doesn't always go well.

REDDY: Right. So what we're trying to do with Candid is basically, you know - kind of explore the idea of using artificial intelligence and machine learning to moderate the community. Now, all of these other platforms - they were founded a couple of years ago. Over the last two years, there's been a lot of advances in natural language processing, in machine learning and kind of artificial intelligence that the machine can understand what you're saying.

And we're trying to apply some of those techniques on Candid. The algorithm on Candid is a learning algorithm, so as we get more data, we learn more. But the idea is to kind of, like, weed out the bad posts, as I call them.

MCEVERS: And walk us through this artificial intelligence monitoring, I guess, of comments. How would that work like, if I posted something that had threats of violence in it or racist comments in it?

REDDY: So the way that artificial intelligence works is it basically - it parses your sentence. We use a deep learning NLP algorithm which basically looks at what you're saying and decides, A, whether it's positive or negative. So it kind of, like, you know, kind of classifies things as having a negative sentiment or a positive sentiment. It then gives it a score of how kind of strong your statement is.

Let's say you said something about someone or you threatened someone, whether it's self-harm or like bullying or harassment, we look for certain phrases and the context of those phrases. And as we, like, get more data, the algorithm feeds upon itself. So it trains based on more and more data, which we feed to it, and over time it gets better and better at recognizing some of these posts.

MCEVERS: You know, I guess the claim could be made that that's the kind of censorship or maybe in lesser situations sanitation.

REDDY: I mean, I think that's a fair point. I mean, here is the thing between, you know, what is "loud free speech," quote, unquote, right? You know, at some level, you should be able to say what you want to say, but at another level you also want to facilitate, like, what I would say constructive social discussion. And Candid is meant for people who want to have constructive social discussions, but want to have their unfiltered opinions, you know, out there.

There is a kind of a trade-off or a fine line that you need to walk because if you let everything in, you know, the fear is that social discussion stops, and it just becomes a name-calling game.

MCEVERS: The app's just come out recently. What has been the reaction to it so far?

REDDY: The reaction's been, you know, kind of mixed. Some people are very, very, very happy with it - is something that I've heard a lot. Some others feel like that we need to be moderating a lot more than we are right now. There have been some controversial opinions - opinions about race, about Donald Trump, which some people believe need to be taken down. So we will have to get our policies right as to what is allowed and what is not allowed so to speak. But, you know, it's just definitely been a very interesting first week.

MCEVERS: Any surprises?

REDDY: Yes. The biggest one being that I live in San Francisco, and, like everybody here, is a part of the tech community. And on my Facebook feed is very, very pro-Democrat and very, very anti-Republican. My Candid feed is 50 percent Democrat and 50 percent Republicans, so I'm really surprised at the number of people who've come out who are my friends who actually happen to, you know, happen to be pro-Trump.

And I think one of the reasons they're not kind of wising those opinions on social media today is because they're afraid to speak out and say, hey, I am pro-Trump for these reasons, and I'm not pro-Democrat for these other reasons. So that's actually been really surprising to me because I don't hear that even in, you know - in my dining room conversations.

MCEVERS: Bindu Reddy is the founder of the new social network Candid. Thanks so much for your time today.

REDDY: Thank you. Transcript provided by NPR, Copyright NPR.