Fake news is a nasty problem facing our world today. The internet has made it possible for anyone and everyone to share information. Furthermore, in the realm of politics, the cut-throat environment sometimes leads to misinformation. Even the climate emergency is facing its fair share of problems regarding misinformation through climate denial videos on YouTube.
Fortunately, there’s a new machine-learning algorithm that can be used to spot out lies in media stories. Engineers of the University of Waterloo envision it being a useful tool for news and social media companies.
The Waterloo tool works through deep-learning algorithms, which are a kind of machine learning algorithm that processes data via successive layers. It extracts increasingly meaningful and complex information throughout the process.
The researchers were motivated to create this lie-detecting algorithm due to the increase of politically motivated, viral deceptions happening online. The system determines whether claims made in social media posts or news stories are supported by other content on the same subject.
Professor Alexander Wong, a systems design engineering expert at the University of Waterloo, said:
If they are: great, it’s probably a real story. But if most of the other material isn’t supportive, it’s a strong indication you’re dealing with fake news.
The team trained its algorithm with tens of thousands of claims, paired together with stories that either rejected or supported them. Then, they tested it using a dataset made for the 2017 Fake News Challenge. The system managed to achieve a 90% accuracy rate in a field of research known as “stance detection,” which refers to the detection of whether the author of a piece of text is against or in favor of the given target.
When the algorithm was given a single claim pulled from a piece of content to compare with other stories about the same subject, it was able to accurately determine if the claim was or wasn’t supported nine out of ten times.
The Waterloo engineers’ achievement with this system is a critical step towards enforcing the truth. It is the first step towards developing an accurate and fully automated system that can help detect fake news. The team suggests it be used by human fact-checkers like an instrument to find the truth.
Wong said:
It augments their capabilities and flags information that doesn’t look quite right for verification. It isn’t designed to replace people but to help them fact-check faster and more reliably.
Graduate student Chris Dulhanty, leader of the project, added:
We need to empower journalists to uncover the truth and keep us informed. This represents one effort in a larger body of work to mitigate the spread of disinformation.
Social media platforms have been coming under much scrutiny by activist organizations lately to crack down on appropriate content. Tools like this are a great benefit to the world because fake news can be a threat to democracy, the environment, and so much more.
