BROADCASTS.com

Professor Hal Abelson of MIT continues to explain the world of machine learning, including its limitations. Have you ever considered that a computer can recognize some images better than people can, but the computer might not recognize that an image is a dog if the picture is upside down? Even the best computer scientists are still working to develop clear and fulsome explanations of why and how machine learning works. As we also discuss with Hal, even the best-trained machines lack what we, as humans, think of as judgment, and machines certainly lack human values. Perhaps new advances will move computers more towards possessing judgment and values – the ability to have compassion or empathy, for example. If so, that could have big implications for how key decisions that affect our lives get made.