Google, Facebook test new fake news algorithm

fake news algorithm

The fake news phenomenon has gathered momentum over the past few years and reached a point at which, some say, it has the power to sway elections. 

Not many elections are bigger than the US presidential contest, in which Republican Donald Trump and Democrat Hillary Clinton were involved in one of the most intense political encounters in recent history.

Not only were both parties engaging in what might be called the traditional mud-slinging, but the campaign was made even more difficult for voters by the outrageous lies and conspiracy theories of what has come to be known as the “fake news” media.

While it could be argued that journalists have historically been somewhat lax at times with the facts, there is a human intelligence behind the stories and those humans responsible can be questioned about it, in court if necessary.

But this new fake news is different: its lies are often outrageous in the extreme, and it employs new techniques such as edited photographs showing scenes that never happened in real life.

And dealing with it using artificial intelligence or algorithms is a challenge which is currently beyond most if not all tech companies.

The other, possibly even weirder practice is to decry a historical event as being fake even when the vast majority of people accept it as the truth, and there were many witnesses.

For example, there seems to be a growing perception that the 1969 moon landings, led by Neil Armstrong, were nothing more than a Hollywood-style film shoot – that the moon landing never happened and the whole moon landing thing was just all filmed in a studio.

These apparently preposterous claims have found their equal – or inverse mirror image – in the newly widespread phenomenon of “photoshopped” false images, often created out of several photographs and illustrations and fashioned into scenes that never happened in real life.

While some might say these are just fun images and most people know they are fake, the problem is that such false pictures and accompanying baseless stories often get listed by Facebook, Google, Twitter and other giant social media outlets as real, factual news – when they are not and should clearly not be listed as such.

Some may argue that people will know they’re fake as soon as they see it, so it doesn’t matter if they’re listed categorised as news or not, but others argue that fake stories did indeed swing the election in Trump’s favour. So fake news does have serious consequences. Fake news and Russian spies – allegedly.

Google and Facebook test new algorithm designed to catch ‘fake news’ Click To Tweet

To combat fake news, if not Russian spies, two of the social media giants which came in for the most amount of criticism – Facebook and Google – are developing algorithms designed to identify and eliminate them from the list or stream or whatever they call it.

They don’t address the central problem of such companies setting themselves up as – and profiting from being – purveyors of other people’s news content, and not even employing enough humans to select the stories and do the editing. All of this is a very important factor in the demise of print and other forms of journalism. And hence the rise of fake news. In a way, fake news was inevitable.

However, we are where we are, and Facebook and Google have launched an algorithmic system to deal with fake news with the help of French media which is being tested in France.

As reported on CNet.com, the two companies are “taking a stand” against the fake news that they themselves helped to circulate.

The project is called CrossCheck, and is designed to “help the public make sense of what and who to trust in their social media feeds, web searches and general online news consumption in the coming months”, says Google.

Google has done many good things online. Its search engine is by far the most accurate in finding what people are looking for, which is why it basically won the web. Its news service – whatever other news media outlets’ quibbles about it – generally does a good job of algorithmically aggregating stories of interest, and not many fake news stories make it through, if any at all.

It’s possible that Google does have a lot of human employees which help the algorithm out along the way, whether it’s in search or news. So it wasn’t really Google which was criticised for distributing fake news.

It was actually Facebook which received most criticism. Perhaps it was being scapegoated, but in any case, the company has since taken the issue very seriously.

As with many websites, Facebook is hoping its 2 billion or so users help out by flagging fake news stories. To be fair, there is just too much content for any company to deal with, no matter how large.

But it’s lucky for such companies that the Internet was always seen as a bit of frontier, a “Wild West” territory, where the rules of the real world don’t always apply, because if they did, a lot of these companies would be sued to bankruptcy, not just Gawker.