A specialist in computer vision and image forensics, Dartmouth professor Hany Farid is at the leading edge of policing the growing threat of machine-learning malfeasance. Last October, he helped launch a five-year program for DARPA called MediFor (media forensics), which is developing software that can analyze hundreds of thousands of images a day and immediately assess if they’ve been altered—spotting, for example, if any color pixels have been disturbed. “Looking ahead to the midterms and 2020, we’re likely to see videos in which faces are swapped and voices are altered so that you get candidates saying something controversial,” he says.

Farid also coauthored a bombshell study in January that revealed how a popular AI software—commonly used in sentencing and parole decisions to predict criminal offenders’ recidivism rates—was no more accurate than a regular person without any judicial experience.

Business Professor, Science & Technology, Internet & Technology Industries, Technology, Communication, Information Technology (IT), Disruptive Technology, Technology Policy, Futurists, Artificial Intelligence Education Professionals, Professors, Science and Technology Professionals, Higher Education, Cybersecurity,