Does "no" mean no if she is a sex robot? Assessments of sexual consent violation impermissibility in the case of sex robots.
Sexual assault is a prevalent issue ruining the lives of many people. As technological developments continue to arrive in the sex industry, we need to ensure that innovation does not propagate problematic activities in society, such as rape. In 2017, Realbotix created a humanoid sex robot, Harmony, with artificial intelligence (AI) enabling the robot to produce natural-sounding speech, to learn and remember things that people tell her, and to exhibit patterns of communication indicative of personality. With the production and distribution of such robots, there come pressing challenges that can be addressed through the study of human perceptions of AI-powered humanoid robots. One of these questions is how to address sex robots’ potential to propagate the normativity of non-consensual sexual interactions between humans. Current sex robots are marketed as ever-willing and never-rejecting women-like entities which can very plausibly be used to enact rape fantasies. It is possible that the permissibility of problematic sexual activities with these advancing humanoid robots could propagate the normalization of problematic sexual activities among humans. It is, however, important to first assess whether non-consensual sexual interactions with humanoid sex robots are perceived as impermissible, and whether these sex robots are perceived to have human-like capacities. The effects that sex robots’ presence can have on the normalization of non-consensual acts will be demonstrated to be of particular concern to the extent that people actually perceive sex robots as possessing some type of human-like mind and moral standing.
In my research, I investigated whether certain features of the sex robots’ AI and embodiment would lead people to perceive the robots as moral patients, as manifested in participants’ judgments of impermissibility of the sexual consent violation of the robots. Using a dyadic model of mind attribution, I found that robots represented as higher in agency and robots represented as higher in experience both elicited stronger judgments of impermissibility of sexual consent violation, when compared to robots who were represented in purely mechanistic terms. This effect was fully mediated by attributions of the robots' sexual consent capacity and vulnerability to harm.
As the results of this study suggest, both agentic and experiential representations of the robots lead to similar mind and moral status attributions. Therefore, we ought to consider regulations on the design and distribution of robots which prioritize both the AI technology (granting robots features such as memory and natural-sounding speech) and the embodiment technology (granting robots features such as bodily feelings and highly realistic human-like bodies). Since robots are granted more moral patiency as soon as they begin to show either one of the two dimensions of mind, both of the technological developments are likely to increase the anthropomorphization of robots and to potentially have an effect on how people view sexual assault in human-human sex. It is reasonable to think that sexual consent violation in the case of sex robots would not elicit such strong moral judgments as sexual consent violation in the case of humans. Still, these findings show that people do not perceive sex robots as fully mindless sex toys once the robots begin to exhibit human-like faculties. Therefore, it might be a matter of technological advancement and the normalization of robotics for people to start perceiving these entities as even more human. Should this be the consequence of technological progress, the worry of normalization of problematic sexual practices will become more pressing. It is likely that we are still at the time when we can divert potential harms by making important ethical decisions such as what kinds of activities we see as acceptable when directed at AI-powered humanoid robots as they are gaining more human-like faculties.