Google has just introduced its new fact-checking feature which lets you check yourself if a news story is true or not – starting in the United States and the United Kingdom.
Google is already sorting news articles by ranking them by relevance. By using specific algorithms Google gives certain news sources preference in its search results. This makes, of course, a significant percentage of total internet traffic. So far Google hasn’t done anything so far to sort out false news stories. Google also claims to have never made any political judgment in their search algorithms. When talking about false news the first thing that is coming to your mind is probably your Facebook News Feed. Facebook is a well-known candidate when it comes to the propagation of false news – especially those of right-wing parties.
Earlier this year former Facebook employees admitted that Facebook “routinely suppressed news stories of interest to conservative readers”. Facebook employees reported that they were instructed to select and influence stories to include them in the trending news module.
“In other words, Facebook’s news section operates like a traditional newsroom, reflecting the biases of its workers and the institutional imperatives of the corporation.” (Michael Nunez, Gizmodo)
Facebook’s algorithm is showing Facebook’s curator team which trends are trending. The curator team then summarizes the news and finds an appropriate headline. After that, the trending news is published to Facebook by linking to selected news outlets. Facebook’s curator team thus has an enormous influence on the global opinion formation of its users.
But intervening and selecting certain news outlets and stories over others does not have to be negative by default. Curators have the chance to block biased news outlets and prefer more neutral news outlets instead.
Facebook is filtering false news and propaganda manually with a team of curators. Now it was simply a question of time when and how Facebook or Google would implement a fact-checking feature for news. Now Google made the first step and it announced a fact-checking feature just one month in front of the U.S. elections.
Google Fact Check
Google announced on Friday that it is adding fact-checking results to Google News. Right below a certain news story, you will find links to different unbiased stories and fact checking analysis.
This is how it looks like today (October 16, 2016):
So Google is right now linking right below of the top Google News article links to web pages where (hopefully) unbiased authors have check proofed the statements made in the article. Now it is basically your turn to read this fact check article to form your own opinion on the trending topic.
Fact checking is done by more than 100 independent fact checking sites who curate together “many thousands of fact checks a year” (Duke University Reporter’s Lab). Bill Adair is the founder of Politifact and he expressed this step as follows:
A huge step for fact-checking: Google News now noting fact-checks in search results pic.twitter.com/kZEQnffouP
— Bill Adair (@BillAdairDuke) October 13, 2016
The problem? Fact Check articles are usually longer than the actual news story itself and algorithms seem not to work to filter out fake stories. Furthermore, fact checks are not available for every story. So how will the future of fact check look like?
How to evaluate Trustworthiness
Today your Facebook timeline is flooded with news posts, news videos, and right- and left-wing propaganda. How can we use today’s technology to evaluate the trustworthiness of headlines, Facebook posts, Tweets, and videos?
We need to look at the cases Facebook and Microsoft dealt with in the past: Propaganda and false stories trending on Facebook Trends, or an AI chat bot who became a Nazi within hours. Those cases make us clear that current algorithms and even bots powered with artificial intelligence have still problems figuring out what is false or right. What can we do about it?
Firstly, I am a strong opponent of the idea that algorithms are determining which news story we get to see. And Google News, Facebook, and other personalized news streams are doing exactly this. If people rely solely on a specific platform it is extremely easy to manipulate them. The company of your choice might simply change its algorithm and you won’t see any news stories of e.g. a certain protest movement.
Instead of blocking content we could calculate an open-sourced index which might indicate if a certain news story is trustworthy or not. How could it work?
Today’s technology makes it possible to check the location of people, crawl unbelievable big amounts of data, and evaluate the quality of articles. By using the massive amount of data available and by using smart algorithms who are able to detect the quality and even the validity of texts we can calculate an index of trust.
By taking into account different factors we could come up with an index from 0 to 100 indicating the probability of trustworthiness:
- Location of the author while writing, recording, researching for an article
- Differences between articles of different sources (are they reproduced or individually researched)
- Writing patterns: Does the author lie or make the story up?
- Number of similar stories around the globe
- Eyewitness reports on social media
- Sources and quotation used
- Translation mistakes
- And much more
This kind of index will make it harder for false stories to spread quickly.
What do you think, should we trust algorithms or humans to filter the massive amount of news data on the internet for us? Leave a comment and I’d love to join the discussion.