When robots were simpler it might have been reasonable to talk about them with generalizations about "people," "needs" and "experiences." But now robots are increasingly involved in day-to-day life, where these broad terms can’t capture all the variety of human experience with the machines. You have to ask: which people? whose needs? whose experiences? When they didn’t ask those kinds of questions, they got (to name a few examples) robots that didn't recognize Black faces, robots that looked like stereotypically female for no functional reason, and robots that don't work for people in wheelchairs.
Katie Winkle is an engineer, designer and social scientist who wants to understand how machines can be designed to benefit everyone -- not only to prevent the effects of bias, but perhaps to get people past some of their biases. (The robot in the podcast illustration is telling the human he’s an idiot as part of an experiment Winkle did in her quest for a robot that won’t just politely ignore sexist comments.) She has been working for some time now on connecting the ideas and insights of feminism with robotics. So of course we spoke about how a feminist robotics would make better robots for everyone
We talked about many sometimes surprising things that happen when robot design is opened up -- for instance, the difference between "participatory design" (when future users are involved in planning a robot) and "participation washing" (which is doing the same old engineer-driven robot design but giving it an air of participation); what happens when people you design for don't think they're in the category you put them in ("this is for old people, not a 67-year-old like me!"); robots as authorities versus robots as entertainers ; making robots that people can adapt to their own personalities; how robots could help people be their better selves (or at least not their worst selves) ; who decides what “better” and “worse” are. Among other topics. Give a listen!
As always with the transcript, remember this is AI-generated and may contain a few errors. Please let me know if you spot any.
To Design Better Robots, You Have to Imagine a Better World