In its early days, Twitter (TWTR) was optimistically viewed as a kind of digital pub where you could hang out with friends, meet new people, and shoot the breeze about your interests.
Like most social media in 2019, it’s now seen very differently — it’s an altogether more dangerous beast.
The platform is a toxic place for women, according to Amnesty International. It’s been dubbed a “handmaiden to authoritarianism”. Abuse of politicians on the platform is rife, as is racist abuse of sports people. Twitter CEO Jack Dorsey’s own account was hijacked in August by hackers who posted a message defending Adolf Hitler.
Biz Stone, the self-described “moral compass” of Twitter, is trying to figure out how to fix this mess.
“We have to question everything,” Stone told Yahoo Finance UK. “If we’re going to survive we’re going to have to make changes.”
Stone was one of four co-founders of Twitter in 2006 but left in 2011 to pursue other projects. He returned at CEO Jack Dorsey’s request in 2017, during a low-point for the company.
“It was basically: come back here and make everyone happy again,” Stone said. “Everyone’s miserable, the stock price is at $13, everyone’s leaving, it doesn’t feel special to work at Twitter anymore, users weren’t growing, the company wasn’t growing, the company wasn’t profitable. It was kind of like, what are we doing?”
It’s a question many people outside of the company have often asked over the years. A common gallows humour joke on the platform is to greet any new Twitter update — more often than not, some small tinkering with the design — with the refrain: “Just get rid of the Nazis.”
Can Stone really get Twitter back on the right path?
‘Why does Twitter exist? There was no answer’
When he returned to Twitter, Stone had no title or brief beyond “make Twitter happy again.” He began by coming up with a mission statement — a north star to help guide the company.
“We never articulated a purpose statement for Twitter. Why does Twitter exist? There was no answer,” Stone told Yahoo Finance UK at the OneYoungWorld summit in London last month.
“I answered the question with: We serve the public conversation. That had a much bigger impact on the company than I could possibly have realised because it was so galvanising.
“If we serve the public conversation, then we are responsible for the health of the public conversation. We’re responsible for the content on the platform.”
This is a quietly radical statement coming from a social media executive. Rival Facebook (FB) has spent years denying it is a publisher, wary of the liabilities that would come if they take responsibility for what people post. Stone himself wrote a 2011 blog arguing Twitter was also just a platform and not responsible for what appeared on it.
“We’re like AT&T: we make the phones and the wires, what people say on them, not our problem, not our business,” is how he now recalls that blog.
Since then, his position has changed.
“What everyone wants to hear us say — and what actually is the ethically correct thing to do and say —[is] we are responsible for the content on the platform,” Stone said.
Twitter’s lawyers “hated” his volte-face, he said.
“The argument that finally won the day was: I still believe that freedom of speech is a fundamental human right, I don’t believe that freedom to have that speech amplified by Twitter is a human right,” Stone said. “You have to play by our rules.”
‘We didn’t think’ about mass manipulation
Part of admitting responsibility involved getting more serious about the removal of abusive or misleading tweets. This has always been a weak spot for the company.
“Traditionally, we’ve been a little bit behind,” Stone said. “We just didn’t realise this mass manipulation — we didn’t think of it. I don’t think anyone was really up on this stuff until all of a sudden it was happening and then we had to play catch up.”
Twitter started to use artificial intelligence to weed out abusive content last year and the company said last month that over 50% of abusive tweets are now being removed before users flag them. The system is “working like a charm,” according to Stone.
“It turns out that most of the time, if not 100% of the time, certain behaviours are always tied to certain types of content,” he said. “If there’s an account that signed up today, no profile picture, it’s at mentioning you hundreds of times today, you don’t follow it, it’s probably harassing you.
“We look at hundreds, maybe thousands, of signals per account using machine learning, which means we can scale, scale, scale. It’s definitely the way to go.”
Fake abusive accounts are only part of the problem, however. Many users are equally unhappy with the real people who post hate speech, mis-information, or far right content, often under the guise of respectability.
US President Donald Trump is the most obvious example of a troublesome tweeter. The platform has struggled to get to grips with how to treat Trump, who has in the past used the platform to threaten war on North Korea and make multiple false statements, according to the Poynter Institute’s PolitiFact.
Shortly after Yahoo Finance UK spoke to Stone, Twitter announced a blanket ban on political adverts ahead of the 2020 US election. The move drew plaudits from the press and came in stark contrast to Facebook, which has pledged to even allow lies in adverts.
However, Twitter’s ads have never had as big a political impact as Facebook’s and there’s a question as to whether someone like Trump would even need to advertise on the platform. The US president has 66.4m followers on Twitter and gets retweeted to millions more.
Would Twitter intervene if Trump tweeted lies about an election rival?
“With regards to just regular tweets, we have the rules and the enforcement of the rules,” Stone said. “If they lie in a tweet, I don’t know if that necessarily breaks the Twitter rules. Maybe you lie in a tweet but people see that and they say, you’re a liar, why would you lie? But if you pay us to spread your lie, that’s a different story.
“We constantly have to look at these policy and rules,” he added. One recent addition was a ban on retweeting world leaders who break the platform’s rules.
So far, the most prominent rule the president has fallen foul of is copyright. A viral video he posted about rival Joe Biden and his son Hunter was taken down for infringing the copyright of the band Nickleback.
Deepfakes ‘the new spam’
Twitter is still getting pelters in the press and the share price recently crashed 20% in one day after an earnings missed. But Stone thinks he and his colleagues have made a lot of progress in picking the company back up.
“The company’s turned around — we’re profitable, we have money in the bank, the stock price is up from $13 to around $40,” Stone said. (Since we spoke, the stock has decline and is now nearer $30).
“People are happy, the service is growing, the company is growing. It’s a complete turnaround. I can’t think of another consumer internet company where this kind of turnaround has happened in a year and a half. Usually, it just dies.”
After playing catch up, Stone now wants to get ahead of the next threat to Twitter: deep fakes.
“We’ve just issued our thoughts on synthetic content — deep fakes, misinformation,” he said. “We’re starting out with asking everyone who’s using Twitter what their experience has been, what should we do?”
Stone was appearing at OneYoungWorld to promote his investment in the AI Foundation, a startup building virtual AI avatars of people that can learn their personalities through talking to them. Stone recalled how he spent time talking to a fake Obama avatar on his phone.
The experience has convinced him not only of the power of this technology but also its risk. Stone is talking to the AI Foundation and others in the space to figure out how to deal with deep fakes, which are videos edited to look like someone is saying or doing something they never did.
“We know this is going to be a thing,” Stone said. “It’s like the new spam. It’s like if you were starting an email company today and you didn’t worry about spam. This is happening already and it’s going to happen more.”
‘Should we have the follow number?’
Stone is also thinking deeply about how to improve the overall level of conversation on Twitter.
“What we really want to have happen on Twitter is to have more people have more conversations, real conversations where they listen to each other, they hear each other, and they come to some kind of mutual understanding,” he said. “Not just, you’re wrong, I’m right, you’re wrong, I’m right.”
One idea he is kicking around is getting rid of the follower count on people’s profiles.
“A lot of this comes down to: what do you incentivise people to do on Twitter?” he said. “Without really realising it, what we incentivised people to do was make that number go up — I want more followers.
“This comes all the way down to a little thing like: should we have that number? Should we have the follow number there? Or should we instead have some kind of indication of the success of this person in terms of conversations?”
Twitter last year launched a project to measure the ‘health’ of conversations on the platform, in partnerships with academics from around the world. The findings could inform how Twitter looks and works in future.
“You know how when you create a password, they give you the password strength indicator? I always thought it would be fun as you’re typing a tweet or an email or whatever — um, this is red, you’re red right now,” Stone said. “You’ve been tweeting really fast, a lot. It gets a little greener if you go back and use a different adjective.
“We’re not going to do that, but it’s that kind of thinking. How do we incentivise people to be more civil and to have more empathy and to have a discussion that is productive? Those are difficult questions. Not just for Twitter, the world in general.”