Is That Self-Driving Car a Boy or a Girl?

As new technologies take on increasingly humanlike qualities, there’s been a push to make them genderless. Apple’s Siri digital assistant unveiled a gender-neutral option last year, and more gender-neutral interactive tools, such as Q and Replika, have recently entered the market. When asked about their gender identities, the AI chatbots ChatGPT and Google Bard each reply, “I do not have a gender.”

The United Nations and World Economic Forum have cautioned against gendering technology, arguing that doing so reinforces societal stereotypes. That is likely true, says Ashley Martin, an associate professor of organizational behavior at Stanford Graduate School of Business. “People are stereotyping their gendered objects in very traditional ways,” she says.

Removing gender from the picture altogether seems like a simple way to fix this. Yet as Martin has found in her work, gender is one of the fundamental ways people form connections with objects, particularly those designed to evoke human characteristics.

In her latest research, conducted with Malia Mason of Columbia University, Martin looks at how people respond when real or imagined products are assigned a gender — or no gender. One of their studies analyzed Amazon.com reviews to see how shoppers reacted to the anthropomorphizing and gendering of robotic vacuums. In two other studies, participants were asked to rate their attachment to male, female, and genderless versions of a digital voice assistant and a self-driving car known as “Miuu.”

Throughout the experiments, Martin and Mason found that gender increased users’ feelings of attachment to these devices –– and their interest in purchasing them. For example, participants said they would be less likely to buy a genderless voice assistant than versions with male or female voices. These findings suggest that gendering a product could lead to higher sales or more consumer loyalty. “Gender facilitates owner-attachment to devices, an outcome that companies spend billions annually chasing,” Martin and Mason write.

Woman or Machine?

While accentuating gender may be good marketing, it may also reinforce outdated or harmful ideas about power and identity. “By highlighting gender stereotypes we’re likely reinforcing men’s greater power in society,” Martin says. The stereotypes commonly associated with men, such as competitiveness and dominance, are more valued than those associated with women. “In leadership domains, those are the qualities that we see and we want in our leaders, whereas women we see as kind and warm and affectionate,” Martin says. “Great qualities, but not as highly valued in leadership.”

These qualities, in turn, are mapped onto products that have been assigned a gender. In her studies, “people describe male and female voice assistants completely differently,” Martin says. “The male voice assistants ‘got things done’ and ‘found the information quickly,’ whereas female voice assistants were described as ‘helpful.’” Assigning a gender to an ungendered tool like ChatGPT would undoubtedly change the way people perceive and interact with it, Martin says.

Yet Martin’s study also found that creating a gender-neutral object was difficult. For instance, if an object’s name was meant to sound gender-neutral, like Miuu, participants would still ascribe a gender to it — they would assume Miuu was a “he” or “she.” “Even if I took away gendered information, people still use gender,” Martin said. Similarly, the fact that so many people have asked ChatGPT about its gender reinforces how meaningful gender is to people.

Marin sees a silver lining, however: She believes that anthropomorphism “provides an opportunity to reduce or reverse stereotypes.” When women are put into positions of leadership like running companies or teaching STEM courses, it reduces negative stereotypes about women. Similarly, anthropomorphized products could be created to take on stereotype-inconsistent roles –– a male robot that assists with nursing or a female robot that helps do calculations, for instance.

Martin hopes that future research will explore how to increase people’s attachment to objects without relying on gender norms. Some of that work is already being done by shifting ideas about gender stereotypes and gender identity. “We’re moving toward a place where society is removing gender more and more,” Martin says, “and whether those strategies are going to remove it from cognition or at least lessen the role it plays in cognition over time is an open question –– perhaps they will.”