Probably exactly what they feel when they are turned on, which is to say, nothing at all. At least that’s what the majority of robotics experts believe, even if there is repeated talk of “empathic machines.” Artificial intelligence ethicist Joana Bryson considers it deceptive and manipulating when some companies or researchers claim that robots can “feel” sadness, fear or pain thanks to corresponding sensors. However, anthropomorphic machines that visually mimic humans are getting better at simulating emotions—and in reading emotions by analyzing facial expressions.
Scientists at the Technical University Darmstadt, for example, are using the robot Elenoide, who seems to nod in understanding, to see how humans react to robot coworkers in offices. Elenoide confirms findings from other studies that show we tend to treat humanlike robots like people—sometimes even treating them better than our human colleagues. Martina Mara, a professor of robot psychology at the Linz Institute for Technology in Austria, cautions that “education and transparency about what the machines can actually do” is what is needed instead of “misleading consumers for marketing reasons” into thinking that robots have feelings like we do.
