Tech entrepreneurs are no strangers to disruption. However, over the past 18 months startups have found themselves facing challenges to the status quo of a more negative kind in the form of fake news. No matter how you look at it, false stories circulating online have become a huge burden for democracies and companies around the globe. “And I guarantee you that it’s getting worse,” says Dhruv Ghulati, co-founder and CEO of Factmata, the startup using AI to detect falsehoods online. Fortunately, he and other entrepreneurs are now trying to turn the tide in the fight against untruths on the web.
Fake news stories are in themselves nothing new. “The business of disinformation is as old as the world,” says James Chappell, co-founder and CTO of Digital Shadows, the digital-risk-management scaleup. Indeed, look back through the ages and you’ll find no shortage of individuals spreading lies for personal gain or to undermine others. For instance, between 1678 and 1681, England fell victim to the so-called Popish Plot where protestant clergyman Titus Oates convinced the country that Catholics plotted to assassinate King Charles II. Despite being an utter fabrication, the lies played on the anti-Catholic sentiments of the time and resulted in the executions of 22 innocent men before the hysteria ended and Oates was sentenced for perjury. “And we’ve got examples dating as far back as Roman times,” says Chappell.
However, long and murky as this history may be, it’s hardly a secret that the dawn of the digital age has exacerbated the situation. Far from being confined to a corner of the internet predominantly populated by trolls, falsehoods on the web have now gone viral. “I didn’t think it was something that affected me,” says Lyric Jain, CEO and founder of Logically, the intelligent-newsfeed startup. “But I started taking it seriously when I was fooled by that meme about Donald Trump thinking Republican voters were stupid.” The meme in question was one of the most well-circulated ones during the 2016 US presidential election. Although, it was neither the only nor the last one. In the past 18 months fake news – often with Russian origin – has seemingly attempted to influence the outcomes of elections in the UK, France and essentially every other western democracy having a vote during this period. As the problem has become critical, people have realised that the consequences of it can be disastrous. “It can deeply affect how our democracy works,” says Jain. No wonder then that Collins Dictionary named ‘fake news’ the word of the year in 2017.
And you don’t have to look far to see how tech has made the situation worse. “The speed of tweeting means you have the ability to reach millions and millions of people instantly,” says Jain. Moreover, the emergence of bots and hyper-targeting techniques means false stories can more easily reach the people they’ll affect the most. “That’s probably the reason why all of this is so damaging,” Jain says.
Making matters worse: people like fake news. In March, researchers from Massachusetts Institute of Technology tracked both fake and true news stories shared on Twitter between 2016 and 2017. Alarmingly, of the 126,000 stories selected, the top 1% of false news stories reached between 1,000 and 100,000 people whereas the truth rarely reached more than 1,000. And they weren’t spread by bots but by individuals who actively shared them. “People love hearing that they are right,” says Jain.
But the times they are a-changing. In the immediate aftermath of Trump’s victory Mark Zuckerberg scoffed that it was “a pretty crazy idea” to think that fake news on Facebook played a key role in the election. He’s since changed his tune, pledging in January to do more to tackle lies on the site. Twitter’s co-founder and CEO Jack Dorsey has made similar pledges and these two tech titans are hardly alone. “There have been significant calls from not only within these organisations but also from regulatory bodies to try to deal with the credibility of the content people are consuming,” says Jain. For instance, in November the EU budgeted €1.1m to its East StratCom Task Force to tackle Russian fake news, which was the first time it had received money from the EU budget rather than being funded by member states’ voluntary contributions. Similarly, Theresa May pledged in January to set up a unit to tackle fake news. Moreover, industry bodies like Newsworks and the European Association of Communications Agencies have called for more discussions around fake news. It was also a hotly debated topic at this year’s World Economic Forum meeting in Davos. Even the pope has taken a stand, saying: “Fake news is a sign of intolerant and hypersensitive attitudes and leads only to the spread of arrogance and hatred.”
And as the problem has become more widely acknowledged, fighting fake news has proven to be a fertile ground for startups. “Technology has transformed our lives,” says Chappell. “Our lives are now lived online and that brings new opportunities but also new risks and people are responding to these dangers. We are seeing a natural evolution of people trying to minimise the negative impact and maximise the positives.” And the need for these startups is only set to increase. “I think we’ll see the role digital media plays in society further grow in importance and therefore the value of modifying online conversations will continue to rise,” says Chappell.
However, it’s difficult to classify exactly what this sector is. “It actually involves a number of different disciplines,” says Chappell. “In that sense it’s tricky to sort of define it as one thing.” For instance, Matter, the media accelerator, has announced that it’s actively looking for media startups to help them fight news and cybersecurity startups like PerimeterX use AI to track and block bots from sites. Moreover, there’s a slew of initiatives aiming to educate people about being more sceptical online and how to get out of their own bubbles. “So I would stop short of calling fake news a sector,” says Chappell. “I’d say that media is the sector and that there are companies that help media organisations combat fake news.”
Given there are lots of different startups fighting fake news, there are also many different ways they go about doing it. For instance, some aim to tackle the huge challenge of fighting falsehoods on social-media platforms. “The scale at which these campaigns occur makes it really hard for social-media companies to respond quickly,” says Chappell. As an example, in a bid to tackle this Facebook has announced it will hire 10,000 more people to work on the site’s security and safety. However, pumping up their staff numbers may not be enough. “If you think about how much content they have to look at, the scale of it is so big that you can’t economically deploy humans to look at every single post,” says Chappell. “You have to automate it.”
And creating that automation is just what some startups are working on. But separating lies from truth is harder than it might look. “What is fake to you may not be fake to someone else,” says Ghulati. Therein lies the problem. While it may be easy for AI to spot inaccuracies about hard data like the population of a country, it’s extremely difficult to create software that can say whether or not nuanced information is true or false. Remember that people still argue about the truthfulness of the Leave campaign’s bus that claimed that leaving the EU would free up £350m to the NHS per week. Even though there is promising technology out there, accurate fact-checking still requires lots of human input. “Fact-checking in terms of automation is a true artificial intelligence – if a machine could do that we would have proper natural language understanding,” says Ghulati. But that technology is still years off into the future.
Recognising the difficulties of attacking fake news once it’s begun to spread on social-media platforms, it’s hardly surprising that many startups are working to take down the sites publishing fake news before it goes viral. Often this means combatting so-called impersonation. “Pretty much all businesses and organisations have some level of impersonation,” says Chappell. One of the most well-known examples happened during the French presidential election when someone set up a bogus website of the Belgian newspaper Le Soir which had a fake story claiming that Saudi Arabia was financing Emmanuel Macron’s campaign. While this example saw someone set up a fake site, purveyors of fallacious stories also find ways to hide their stories among real media outlets’ pages or on real corporate sites. “What we do when we detect this is to tell the customer how the criminal has put it in place and the impact of it on their organisation so they can brief their customer-service agent and their personnel,” says Chappell. “We also operate a take-down service.” By nipping it in the bud, startups like Digital Shadows hope to prevent the proliferation of phoney content.
Despite these efforts, the fight against fake news won’t be over anytime soon. “We’re only just seeing the start of this really,” says Chappell. “It’s still very early days. The attackers will continue to evolve their techniques and innovate but so will the defenders.” Because of this simultaneous evolution, it’s impossible to see one side or another win the war. “I think it will be more than a series of battles,” he concludes. But at least, with these startups backing them, the champions of truth may have a fighting chance.