The past 12 months have seen a surge in what is often referred to as Fake News. In reality, the majority of this is nothing more than an inability to check facts accurately.
We have seen this have a huge impact on the world over the last year. The claim that the UK spends £350 million per week to be a member of the EU, for example, and Donald Trump’s claim that 3-5 million illegal immigrants voting for Hillary Clinton led him to lose the popular vote in the 2016 US election, in particular spring to mind. These numbers are both easily debunked with a 1 minute Google search. The £350 million figure failed to take into account rebates or money that comes back to the UK from the EU and was strongly criticized by the UK Statistics Authority, who said that its use was ‘potentially misleading’ and was ‘disappointed’ that it continued to be spread. Donald Trump’s 3 million votes claim was also thoroughly debunked given that a national investigation by the Carnegie Corporation of New York and the John S. and James L. Knight Foundation found only 56 cases of non-citizens voting in all elections between 2000-2011.
Given how simple it is for these claims to be debunked, it is surprising that these facts still persevere amongst many in society. According to a survey by Morning Consult and Politico, 47% of all those who voted for Donald Trump believe that he won the popular vote, representing 29.6 million people or around 9% of all Americans. In the UK, an Ipsos MORI poll shortly before the EU referendum found that 47% of voters believed the NHS, despite the well-publicized criticism of the figure from the UK Statistics Authority and many moderate media sources. In fact, only 39% of people believed it to be a false figure.
One of the key reasons these untruths become so entrenched is their reporting in the media, who either deliberately misinformed their readers and viewers, or did not adequately check the veracity of their sources. Sloppy mistakes are understandable to a degree. There has been a drop of 42% in the number of journalists in the US since 1990, yet the move from print to online means the amount of content they must create has increased, meaning they have far more work to do.
However, rapid advancements in deep learning and AI in recent years may solve the problem for overworked journalists by helping to ensure the veracity of their sources. The problem is particularly pronounced among citizen journalists, who have become increasingly prevalent in recent years as the internet and social media allows figures to publish anything in an unregulated market. Their articles are often poorly sourced and contain incorrect information, which manages to make make its way into the public domain with little or no consequences. With AI and deep learning technologies to verify what they’re writing, this increasingly influential medium could become better respected and more informed.
Louise Mensch and Claude Taylor, two anti-Trump ‘citizen journalists’ have been very public victims of this, having been conned by a hoaxer who claimed to work for the New York attorney general. They believed this source, apparently without checking backgrounds or facts, before publishing salacious rumors about Donald Trump being investigated for multiple crimes that had no basis in reality.
Talking to the Guardian, the source of the hoax claimed ‘[Claude] Taylor asked no questions to verify my identity, did no vetting whatsoever, sought no confirmation from a second source – but instead asked leading questions to support his various theories, asking me to verify them.’ The records of many of these hoaxes would have been easy to find, even a rudimentary internet search would have picked up on many of these issues. This is something that AI could do incredibly easily if used correctly - journalists would simply need to ask questions of it and then pull up information on the subject. Through creating a simple report on the source, it would save hours of researching hundreds of pages to understand the veracity and trustworthiness of those they are dealing with, which could also be shown to readers so that they understand the content is genuine.
There are already some rudimentary fact checking technologies available that can evaluate information about social media accounts, such as BotOrNot, Botometer, and Twitter Audit, who are all using basic techniques to investigate whether Twitter accounts are real people or bots, something that is increasingly difficult to work out. Donald Trump, for instance, has even been caught out several times by accounts who have been bots, which shows that this is a problem that almost anybody can be conned by.
Using AI and deep learning to create a quick report on the statistics and quotes used should be relatively simple, scanning through complex and variable data sources to discover patterns that show whether information is correct or not. These tools could even check that images accompanying the article show the correct picture and context. A recent article on Breitbart, for instance, could see them end up in court after they used an image of Lukas Podolski, a German soccer player who has appeared for his country 130 times. The image of Podolski and another man appeared under the headline ’Spanish police crack gang moving migrants on jet skis’, but a 10 second Google image search would have shown that this was actually an image of Podolski on a Jet Ski trip during the Rio 2016 World Cup. A simple AI system would have picked this up almost instantly through image recognition and allowed them to avoid the embarrassment and potential law suit that Podolski is reportedly considering against them.
With the amount that needs to be covered by journalists, a tool that can replace the lengthy manual fact-checking element of their words, figures, and images could allow for increased productivity, more accuracy in the stories written and, most importantly, would stop these untruths having such a huge impact on our society.