Computer science is definitely reaching the danger zone when actual words can be spotted using MRI scans. In the image above, researchers from the Machine Learning Department at Carnegie Mellon University in Pittsburgh predict what the words “celery” and “airplane” look like when someone thinks of them, and then they compare the prediction to actual brain scans, with frightening similarity. The study was “calibrated” with nine students, each thinking of 58 different words. Tom Mitchell, one of the lead researchers, told Reuters the goal is to determine how the brain organizes information, but how do we know Dr. Mitchell won’t abuse this newfound power by, say, winning a billion dollars on Jeopardy? We don’t, is my point. [Reuters]
