×

Will clamp down on malicious automation by bad faith actors: Twitter VP

Twitter VP Colin Crowell | Sanjay Ahlawat

Twitter has revolutionised the way people consume and share news. It has become one of the most potent public conversation spaces between people in power and the person on the ground. It has become a handy tool for the politician to speak personally to his constituency, or even his rival. As India prepares for the mother of all battles—the Lok Sabha elections—THE WEEK caught up with Colin Crowell, Twitter's global vice president for public policy, government, and corporate philanthropy efforts. Crowell opened up about Twitter's preparations to deal with the malicious content on the platform, elections and growth in the Indian market.

Excerpts from the interview.

Indian government has requested you to check the misuse of Twitter, particularly for the 2019 elections. What steps are you taking?

We have conversations with governments in the countries where elections are held, to know how social media can be used positively. What can we guard against? Going back to the 2016 elections in the United States, fairly substantiated allegations of foreign interference were made. What we learned there, we will bring to the subsequent elections. We had the benefit of elections in other parts of the world—including Mexico, Brazil and midterms in the US. One vulnerability we identified was the use of malicious automation by bad faith actors who wanted to distribute certain kinds of content.

We have invested heavily in technology to counter some of the problems. We have invested in machine learning to look into the behavioural characteristics of accounts. People try to misuse our openness. We have to be alive to such possibilities.

The same openness is also one of Twitter's great strengths. Openness allows people to see what is happening in real time. Journalists and media organisations can see things and provide a real time corrective to the issue they see manifesting on the platform.

Does this mean that anytime anyone is using automated bots to send malicious content, you will be able to block it?

We have invested significantly and made significant progress. We think the fight against spamming will be a struggle till the end of time. Spam has been there since the beginning of the internet. But, with the counter-measures we bring, we hope to diminish those prospects.

Proliferation of abuse and fake news has led to unsavoury incidents. How do you check such content?

If it is not automated, we receive reports from average users, as any user can report directly. We also receive reports from safety partners and NGO partners in the market who have been trained in our rules. We have invested in human resources so that we have 24x7 coverage around the world. A lot of times, those reports are highly contextual. You have to understand the nature of the conversation that's happening. You have to understand the language set—the nuances, the slang that may be involved. But we have very strong rules against abusive gender-based harassment and threats of violence on the platform. In the last couple of years, we've prioritised safety on the platform in a way that has resulted in some 30 different changes in policy and product, which has led to a 10-fold increase in the number of safety actions, as opposed to a year ago. We are in no way saying that the work is anywhere close to finished. But we do think that we are cognisant of the problem and we are working diligently to make sure that we get our hands around it.

What arrangements or teams do you have in India to understand contextual issues?

Sometimes, things are very contextual. We do have teams all over the world, and we do have people in India. We have people who are skilled in Indic languages. We have taken a note of the fact that if you compare previous elections to the most recent state elections in India, the top trending hashtags were not in English, but in Hindi. It is a change, and we know that as Twitter grows in India, it is growing in language sets other than English.

What more such steps are you taking for the upcoming Lok Sabha election?

So, one thing that we are going to implement in India is something that we test drove most recently in the United States. We are going to develop, more completely, an ads transparency centre which we will deploy so that when candidates and party are engaging in electioneering ads on Twitter, those promoted tweets will have an indicator stipulating that it is promoted political content and who has paid for it. Then, you can click on that. It will take you to a transparency dashboard, which will give you information about the advertisers, the tweets they are advertising, and how much they are spending. So, we think that the additional information for people using Twitter in the market will give them a window into how the election is unfolding. But it will also, I think, provide a resource for election authorities, journalists and others who are covering the races.

What kind of projects are you handling with the Indian government?

We have a range of interactions with a couple of government departments. There is a digital divide in Indian internet usage, including on Twitter. We wanted to showcase and celebrate some lesser-known women, which we discussed with the ministry of women and child welfare. We have relationships with the police departments and the ministry of external affairs, with respect to the use of Twitter in discharge of responsibilities. In the pilot phase of Twitter Sewa, we had an arrangement with around 10 government departments. We are seeing how they have performed. So, when the next government comes in, there could be further arrangements.

What is more beneficial for the Twitter as an organisation—the sheer large number of people joining it, or the conversations they are having?

The overarching mission of the company is to serve the public conversation. More interesting for us is how people engage on the platform in back and forth and conversational tweets. It also goes to the nature of what makes Twitter special. We are public, we are open, and it can lend itself to being eminently conversational as well. I think our feeling is that people will derive the most benefit out of a public, open and conversational platform, particularly when there is a national conversation to be had. Sometimes those national conversations may be during the World Cup, and following the national team. Sometimes, it may be around the Olympics, or it could be for the Oscars.

But, certainly, in any democracy and certainly in a democracy that is as large and robust as India, we will have a national conversation.

Twitter's strength is its openness. But then, it also raises the question of regulation. It could be self-regulation, or the ones imposed by different governments. How do you deal with it?

We have community standards. They deal with governance, conduct on Twitter. Those rules address safety issues, harassment, threats of violence and impersonation. Obviously, the contours of the line that constitutes free expression will change from society to society, and from culture to culture. The laws around that will also change from place to place. We do have the ability to cater and tailor content issues on a national basis. And we have done that in India in the past. When we do that, we try to provide as much transparency to our users and the public as possible. So, if something is violative of Indian law, or in violation of a French or German law, that is different from what might be in our global rules.

We can cater to that, but when we do that we do provide notice of what official government missive we have received—a court order perhaps—and we put that on the lumendatabase.org which is a joint venture between the Electronic Frontier Foundation and Harvard's Berkman Center. It is a public place for people to see and decide whether or not we made the right call, whether we should have agreed to it or fought it. But, on platform, we don't just make content disappear. So where the tweet was, or where the account was, we give an indicator that the content has been removed in country x pursuant to this law. Click here for more information. But content that violates one country's laws, but not violative of others, would still be visible in the rest of the world.

How many requests have you received from the government of India?

We make that information public every six months in the Twitter Transparency Report. You can see what reports we got from the government of India, or the compliance rate.

Specifically what kind of complaints or issues are flagged in India?

Its hard to characterise them because it could be a range of different things under under different laws.

Twitter is trying to expand into other Indian languages. Can you tell us about that ?

We can attribute Twitter's inexorable march into different parts of the country, to a large extent on the growth of non-English (regional) languages, and particularly the adoption of smartphones. I think there's definitely a correlation between the two there. We have also launched something we call Twitter lite, which is a low data version of the app. We have launched that here in India. So, that will help people who might be price-conscious consumers, concerned about data consumption.

What can you say about growth of Twitter in different regions of the country?

I think it's fair to say that we've seen a growth in non-English languages. The growth has come outside of the major cities, because that would be more reflective of the language sets. The growth is very rapid in south Indian languages.

Anything unique in Twitter usage in India ?

What strikes me is the incredible richness and diversity of what you see there. It is emblematic of the diversity of the country itself. It is really wonderful to see the experience of Twitter being used as a vehicle for the democratisation of access to public officials. The democratisation of access is something that, I think, has been tried in different markets, but doing that in a very local, hyper-localised way is very special. It will unfold through a variety of different political parties, different candidates in different geographies, different language sets between now and the end of the election process. So, I think that will be very special for people to see here and participate.

You see Facebook as a competitor?

Yes, certainly to the extent that we are an advertising-driven platform, and they're an advertising-driven platform. So, we compete for advertising, but people have a tendency to say Google, Facebook and Twitter in the same breath—quote unquote the social media. But the reality is that different platforms have different attributes, and we want to play to our strengths and that's how we think we distinguish ourselves for the people who use Twitter and the people who will be joining Twitter hopefully in the next several months.

What new features or innovation can we look forward to in Twitter?

I've been with Twitter for seven and a half years now. And when I joined Twitter, it was 140 characters and a link. If you clicked on the link, it took you somewhere else. Now, it's 280 characters. It's more global. But, it's not just text. It's also graphic information. It's screenshots. It's pictures and and photos. It is also video. But, importantly, it is live broadcast as well. You can go live from your mobile handset, anywhere where you have a good internet connection. That's incredibly empowering. It allows people to bear witness to history, and everybody has the power to become a broadcaster, so to speak. That's an amazing change in technology, really, in only seven and a half years. The ability of anybody to be a broadcaster and reach the world in moments is revolutionary. That's where Twitter is, just after seven and half years of my personal experience with the company.

TAGS