Working alongside Adobe's researchers, a doctoral student at Oregon State University has managed to come up with a new training method for AI systems that is both more efficient and has a lower likelihood of providing solutions that contain ideas and behaviors that can be considered socially biased.