Teaching Astronomy in the Age of Misinformation

Chris Impey, University of Arizona

It’s a time of unprecedented assault on facts. While most disinformation centers on politics and culture, collateral damage to science is substantial. Recent years have witnessed increasing disagreement about facts and data, a blurring of the line between opinion and fact, and a rising volume of misinformation online. For example, astronomers might be perturbed by over a hundred million web pages devoted to UFOs and astrology. This talk describes research on a novel artificial intelligence system designed to detect science misinformation online. Neural networks are trained using curated sets of legitimate and misleading or “fake” articles. In tests on climate change and evolution, the neural networks achieved 90% accuracy classifying articles as real or fake. This testbed is currently being expanded to astronomy-related topics like astrology and UFOs. The technology will be deployed in two tools aimed at helping students navigate science online. The first is a web browser extension to judge the veracity of a science source in real-time, giving a color-coded Bayesian probability that the article is legitimate, and if it is not, referring users to legitimate sources of information on that topic. The second is a smartphone app that gamifies the technology and lets students classify articles as “real” or “fake,” competing with friends and family in this task. A third tool is the use of the same machine learning to identify “claim-evidence text pairs” in student writing, to aid instructors in teaching the appropriate way to write about a scientific topic.

Share

Teaching Astronomy in the Age of Misinformation