In the coming decade, the problem of people believing more in fake news than true information will get more intense and entrenched. Couple this with the numerous surveys conducted last year highlighting the growing concerns of Americans about the effects of fake news on elections, and we know that the era of 鈥榓lways present but always suspect information鈥� has begun.聽

During the last few years, rumor-mongering on various social media apps has led to several unfortunate events including the loss of lives in India. With more than 120 million of its total population of 210 million using WhatsApp, Brazil, in the recently concluded elections, got a first-hand experience of the real concerns around fake news.

So, how did we get here?

Just seven years ago, global internet users comprised only 29% of the world population. Today, that number has grown to almost half of the world population. Rapid growth in mobile internet access, especially in the developing world, has been a game changer. That, however, is only one side of the story.

With increased access to social media and indiscriminate sharing of our thoughts, opinions and other assorted personal information, we鈥檝e opened ourselves up to companies harvesting our data. The indulgence of some in the alternate reality world has led to them playing along with headlines and news that they might have already known is false. In a Statista survey in 2016, the number of respondents who had knowingly shared 鈥渇ake news鈥� was 14%.

Then there is the monetization opportunity. Increased access and sharing has led to indigenous ways of making advertising dollars. Setting up fake news sites and driving traffic to these sites leads to monetization opportunities, which has allowed trolls to have a field day. 聽

Another lesser discussed reason for the spread of fake news is the 鈥渒nee-jerk reaction syndrome鈥�. The penchant for users to crave for likes / comments / shares of their posts has led to instantaneous reaction to headlines / articles. In their quest to be the first to share or react to sensational news stories, users end up not performing any due diligence about the veracity of the stories. This, oftentimes, when others might have already called out the article or story as fake even in the comments section of the same post!

This raises an interesting question. Why aren鈥檛 users depending upon trusted sites to consume news? This is where bias comes into play. A recent Gallup poll found that 62% of Americans believe news is biased and 44% say it鈥檚 inaccurate. In times when the news media could separate fact from fiction, their own perception in the minds of consumers hinders this process. Add to this the cognitive bias that users themselves exhibit and the result is a potent mess.

Each stakeholder seems to be playing a part in getting fake news to travel further and faster than real news.

What is the way forward?

While there is no cure for cognitive bias, there are thankfully many ways in which technology can help in the identification and verification of fake news. Here are three means to make the maximum impact in the fight against fake news:

• Using bots to find bots

Studies have shown that bots are the most active in the first few moments of an article from a low-credibility site being published. Bots tend to tag celebrities or others who would amplify the reach of such articles. Once the bots have done the initial work, humans typically tend to take over from there and continue the disinformation, thereby playing into the hands of the bots. With the latest technologies at our disposal, we can use bots to find such bots. The easiest way to find such bots:

o聽聽 Look up the IPs / geo-locations of such bots - Analyze the domain / account from which the bots are posting

o聽聽 Sentiment Analysis

o聽聽 Who follows the bots 鈥揗atching the language and the region

• Reward crowdsourced fact checking by using blockchain

Currently, an underutilized medium for fact checking is the user base of the social media platform (for example) itself. While there are bots or humans spreading fake information, often users challenge the article or the post by posting relevant information including links to the correct news in the comments section. Co-opting the users themselves in the battle against fake news will go a long way in arresting the problem and the fact that this approach can scale is even better.

• Amplification of real news

In times of fake news, the antidote would be even more viral spreading of the real news. This is where trusted news organizations would have to take the lead. Because these news organizations have journalists on the ground and digital media teams monitoring the trends on social media and in the real world, they can also team up to post the real news on their platforms. This has the dual benefit of not just overpowering fake news but driving greater engagement towards the respective news sites. There are a variety of channels through which such organizations can amplify the real news (websites, messenger platforms, even by placing links as ads in extreme situations).

A concerted effort by all the stakeholders involved (news organizations, users and technology providers) will go a long way in reducing the ill effects of fake news.

A method to the madness

While the easiest way to detect fake news is to run statements, data and claims through a pre-compiled database, the challenge lies elsewhere. It is not immediately predictable as to which piece of information will be used by which group or individual to target an individual / group / organization. The hallmark of a piece of 鈥渇ake news鈥� is the way information gets twisted to target or support entities. Hence, the detection of a piece of content being potentially false and the ability to identify such content from billions of pieces of content being generated becomes crucial in a scalable model for the detection and verification of fake news.

91快活林 has been working with leading publishers providing trust and safety solutions.聽As part of its trust and safety capabilities, 91快活林 has developed a comprehensive approach toward fake news detection and verification.聽

The key highlights of 91快活林鈥檚 fake news detection and verification approach are as follows:

• Leverage the Trigger-means-Target (TMT) model for instantly filtering potentially fake content from billions of pieces of content
• Build a taxonomy of entities that are frequently the victims of fake news and monitor social media for any potential attempt
• Verify the various types of content including:
  • Textual content
  • Memes (extraction of textual content from images)
  • Fake image detection (was an image edited maliciously?)
  • First occurrence of an image
  • Video content verification

91快活林鈥檚 framework for the detection and verification of fake news is a combination of human and machine intelligence with 90 percent of the process being automated. The process is verifiable, repeatable and scalable.