Fake news has been one of the biggest cultural phenomenon of recent years, and much debate has raged about how we can best detect and remove it from our midst. The first challenge this presents is how to actually define what is and is not fake news. As politicians have argued this point, a team of British researchers have come up with a mathematical model to do just that.
The work, which was described in a recently published paper, aims to simulate the impact fake news has on referendums and elections. The researchers defined fake news as noise with specific properties that underpinned the bias within it.
Maths has actually been good at filtering out noise for some time, with filtering theory used since the 1950s to distinguish between bias and random noise to allow us to get at the underlying signal. This approach is used by the researchers to model how voters interpret the news they consume.
The model revealed that people tend to fall into one of three categories:
- Those who are generally unaware of the existence of fake news. These people are thus confident that their views are well informed.
- Those who are aware of fake news, but believe they can distinguish fake news from authentic news. Their awareness erodes their confidence in their views because they don’t always trust what they consume.
- The third group are those who are expert at spotting fake news and instantly discount it. They’re confident in their views and unaffected by the fake news out there. Whilst this is the ideal, it’s a group the researchers believe is very small in number.
They use these categories to simulate how random news articles influence a fictitious electorate. The simulation was run over 1,000 times to try and get an accurate insight into the influence of fake news on our voting preferences.
False influence
Perhaps unsurprisingly, those in the first group were easily swayed by fake news, whilst equally unsurprisingly, those in the third group were largely unaffected by it. The second group however were more interesting. They proved aware of the fake news, but were not aware of when it was created, so they tended to overcompensate for the potential of fake news infiltrating their news feed.
“One can interpret this as an indication that mere knowledge of the possibility of fake news is already a powerful antidote to its effects,” the researchers say.
Of course, politics is nothing if not a prickly domain, and there is much evidence to suggest a strong form of confirmation bias exists when it comes to our political beliefs. Quite how fake news fits into that narrative for the three groups is not something that the paper covers. For instance, does fake news merely confirm our existing beliefs or actually change them?
Nonetheless, it seems sensible to believe that a lot more needs to be done both to counter the availability of fake news, and to better inform people to be able to detect it. It’s an issue that will no doubt rumble on, and this paper is an interesting addition to the debate.
