Researchers at the University of Michigan have devised an algorithm that does a better job of identifying fake news than humans are capable of doing. The advantage is not huge. The algorithm can identify fake news about 76% of the time versus 70% for humans.
It’s the speed of detection that makes the new tool valuable. Fake news often appears, ricochets around the internet, and is gone before a human has time to evaluate it. The algorithm can do in seconds what might take a human hours to do. By the time a story has been verified or debunked, the horse is out of the barn and the damage is done.
Rada Mihalcea, the computer science and engineering professor spearheading the project, tells Science Daily an automated solution could be an important tool for sites that are struggling to deal with an onslaught of fake news stories created solely to generate clicks or to manipulate public opinion.
The algorithm works by analyzing linguistic characteristics like grammatical structure, word choice, punctuation, and complexity. How it got created is an interesting story in itself. Mihalcea and her team used the services of Amazon Mechanical Turk, an online service that allows people to rent the brains of other people to perform cognitive tasks — like writing fake news stories — for a fee.
No need to hire someone and deal with all the hassles that go with having actual employees. Simply submit your request to Amazon, which will find people to complete the required task. Known colloquially as MTurk, Amazon apparently is not aware that name might be seen by some as a slur against people of Turkish descent. The crowdsourced team members were paid to take actual news stories and rewrite them as fake news items which were then fed into the algorithm so it could learn the difference between the real stories and the manufactured ones.
“You can imagine any number of applications for this on the front or back end of a news or social media site,” Mihalcea says. “It could provide users with an estimate of the trustworthiness of individual stories or a whole news site. Or it could be a first line of defense on the back end of a news site, flagging suspicious stories for further review. A 76 percent success rate leaves a fairly large margin of error, but it can still provide valuable insight when it’s used alongside humans.”
The research and the algorithm are available online to anyone interested in using them. Mihalcea says they could be used by news sites or other entities to build their own fake news detection systems. In the future, such systems could be made more accurate by incorporating metadata from the links and comments associated associated with a particular online news item.
Of course, sophisticated hackers could create their own anti-algorithm algorithms designed to fool the fake news detection systems. In the final analysis, we all have to use our own judgment when assessing the things we read online. People may not be able to cry “Fire!” in a crowded auditorium, but they are free to do so online with impunity.
Since the death of journalism brought on by the digital revolution, what used to be a well-tended garden is now little more than a weed-choked space where the most pernicious concepts are allowed to metastasize unchecked. The most effective fake news detector is the one located between your own ears.
Don't want to miss a cleantech story? Sign up for daily news updates from CleanTechnica on email. Or follow us on Google News!
Have a tip for CleanTechnica, want to advertise, or want to suggest a guest for our CleanTech Talk podcast? Contact us here.