Author: Science News Staff / Source: Science News
Computer programmers are building deception-detecting algorithms to fight the onslaught of fake news, Maria Temming reported in “People are bad at spotting fake news. Can computer programs do better?”(SN: 8/4/18, p. 22).
Reader Lou Floyd found the story compelling and troubling.
“It points [to] a major problem facing us all today that affects the very foundation of our government and society,” he wrote. Automated tools to detect fake news and track how it spreads are intriguing, Floyd wrote, but he was concerned about how often such systems might mislabel true news as false.Scientists are still getting a sense of which factors might indicate an article’s truthfulness, Temming says. “Before deploying any computer program to flag false news, researchers should closely examine how frequently that program throws the baby out with the bathwater.”
Computer programs that err on the side of caution, mislabeling more true news as false, may pose less of a problem if the programs funnel suspicious stories to human fact-checkers, she adds. “This is how many programmers currently envision their algorithms being used,” Temming says.
Reader David Roy shared his August 4 issue of Science News with a family member. Roy wrote: “One of our cats, a 5-month-old … named Z, has a head for science — Science News, actually.”
Floyd also wondered how fake news–detecting technology could keep up with fake news peddlers’ ever-changing methods to avoid detection.
This is a common question, says computer scientist Benjamin Horne of Rensselaer Polytechnic Institute in Troy, N.
Y., who develops…The post Readers focus on fake news, neutrinos, and more appeared first on FeedBox.