Last week, we released a survey that showed many Americans are not very optimistic about the benefits of artificial intelligence (AI) and discussed the implications that this might have on U.S. competitiveness in AI. One notable finding from this survey was that women had less favorable views of AI than men. While more research is necessary to fully understand the reasons for this difference, these findings suggest that government, academia, and industry should do more to address gender diversity in the field and ensure an equitable distribution of the benefits from AI.
As shown in Table 1, we asked Internet users whether they agreed or disagreed with three statements about their optimism for the future of AI, and responses had a clear divide by gender. First, 41 percent of men, compared to only 31 percent of women, agreed with the statement, “Technological innovations, like artificial intelligence and robotics, will make the world a better place.” Second, 36 percent of men, but only 27 percent of women, agreed with the statement, “Technological innovations, like artificial intelligence and robotics, will make workers better off in the future.” Finally, 34 percent of men, compared to only 21 percent of women, agreed with the statement, “Self-driving cars will lead to safer roads and fewer accidents.” These results show a remarkable gender divide, and one that the public and private sectors will need to address if they hope to overcome public resistance to AI.
Table 1: U.S. Internet users’ opinions on artificial intelligence and robotics, by gender.
Our findings align with other research on this topic. For example, a 2019 global survey of consumers by Ford Motor Company found that 44 percent of women agreed with the statement, “I am afraid of artificial intelligence,” compared to 37 percent of men. One possibility for the gender divide is that people fear the unknown. According to the same Ford Motor Company survey, 48 percent of women, compared to only 39 percent of men, agreed with the statement, “I don’t really understand artificial intelligence.” However, it is very possible that men are overestimating their understanding of AI, or women are underestimating theirs. Psychologists have shown that in general people tend to think they are smarter than they really are, but women tend to underestimate their abilities compared to men.
Another possibility for the gender divide is that women do not expect the private sector to ensure they fully benefit from advances in AI. The tech industry has made some noticeable gaffes in developing products which overlook women’s needs. For example, women have complained about popular personal health tracking apps that do not include the ability to track their menstrual cycles or fitness apps or smart scales that do not have a pregnancy mode. And these oversights reflect a broader trend for some companies to build products designed for male consumers, such as phones that women with smaller hands may find difficult to use. These types of slights may be part of the reason why in the latest Edelman Trust Barometer, an annual survey of public opinions about trust in various institutions, there is a 15 point difference between the percent of American women who trust businesses versus men. Richard Edelman, the CEO of the public relations firm bearing his name, has noted that “Women, in our study, are substantially more skeptical about innovation. Substantially.”
Fortunately, many in the tech industry appear receptive to change. For example, in response to feedback, Apple added extensive information about reproductive health to HealthKit, its health tracking app, and Amazon added a “disengage mode” to mode to Alexa, its virtual assistant, that activates in response to sexual harassment. Preventing these types of problems will obviously require companies to not just respond to specific complaints but embrace inclusive innovation earlier in the design process.
One way to do this is by increasing gender diversity among those working in AI—which is at dismal levels. A 2018 study by the World Economic Forum used data from LinkedIn to determine that globally only 22 percent of AI professionals are women. And in 2017, only 13 percent of top AI researchers in the United States were women. Fixing this will require not only recruiting more women to computer science—women account for 18 percent of undergraduate degrees and 30 percent of graduate ones—but also making the field more welcoming to women. Until this past year, one of the top AI conferences was notorious for sexist jokes and innuendo.
The only way the United States will be able to remain a leader in AI is if the government works with the private sector and academia to pursue an ambitious national strategy to keep pace with foreign competitors. To do this, the country will need both widespread public support for AI and access to workers trained in AI across every demographic. Thus increasing inclusiveness and diversity should be at the forefront of national efforts to win the global AI race.
AI is a transformative technology, but it will only achieve its full potential if all people stand to benefit from its development. Ensuring this outcome will require the public and private sectors alike to understand and address the reasons for this gender divide. Thus, increasing inclusiveness and diversity should be at the forefront of national efforts to promote and accelerate AI.
Image credit: Voice of America