Do Gender Stereotypes in Advertising Impact AI?

In recent years, an increasing number of research projects exploring gender issues in advertising have identified potentially harmful effects for individuals and for society associated with portrayals of certain gender stereotypes in advertisements. In 2019, the United Kingdom’s Advertising Standards Authority (ASA) banned advertisements depicting gender stereotypes that are likely to cause harm or serious or widespread offence (Regulatory Statement on Gender Stereotypes in Ads). Additional guidance on the types of negative gender stereotypes to be avoided was also issued (Advertising Guidance) in order to help advertisers reduce the potential for harm caused by portrayals that invite assumptions about the ‘proper’ roles assigned to adults and children that restrict how they see themselves and how others see them in light of their gender. The UK is certainly not the only country to take the risk seriously. Several others have discussed, or are currently discussing, laws that aim to minimize the adverse effects of gender stereotypes in advertising (Martínez Alles, 2019).

Gender stereotypes are widespread beliefs about the roles and specific characteristics that are projected upon people by virtue of their male, female, or other non-binary identity. The empirical evidence strongly shows that advertising influences gender stereotypes by reinforcing, confirming, reproducing, accentuating, and perpetuating them in an ongoing feedback loop. Advertisements usually convey images of what is feminine or masculine and associate them with differentiated expectations placed on the behaviour of women and men, thus influencing how people conceive and pursue their life plans and socialization. Reinforcement in social media can lead audiences to accept that the stereotypes are objective, true, unquestionable, and permanent. In this way, rigid beliefs in society are reproduced and strengthened. In other words, advertising plays an important part in lending beliefs that are socially constructed a natural, unquestionable quality.

Although we cannot exclusively blame advertising for the gender stereotypes entrenched in our society, we cannot deny their impact (even if only subliminal) on the aspirations and expectations of people, whether regarding the careers they aspire to or in society in general. For instance, research shows that the common representation in advertisements of women in kitchens and handling household chores and childcare, rather than in leadership positions in workplace settings or performing creative activities, helps consolidate preconceptions that associate femininity with permanent caretaking, creating a discourse that excludes them from aspirational and leadership roles as well as many other activities that are represented as predominantly masculine.

A timely example is found in the scarcity of women in the broad field of computer science, where the predominance of male workers is well documented. The predisposition to believe that girls are ‘not as good’ as boys in math and science yields a stereotype that is further reinforced by the focus on male ‘tech geniuses’ or ‘giants’. This feedback loop dynamic makes it easier to understand why many girls tend not to pursue careers in STEM (science, technology, engineering, and mathematics). According to the World Economic Forum Global Gender Gap Report 2022, the problematic underrepresentation of women in STEM fields persists despite efforts to rectify it. Looking at the distribution of female university graduates across all fields and comparing it to the distribution of male graduates, ‘the percentage of women graduates in Information and Communication Technologies (ICT) is 1.7%, compared to 8.2% of men graduates. In Engineering and Manufacturing the same figures are 24.6% for men and 6.6% for women.’

This problem creates another for those concerned with ensuring that the increasing number of decisions made by programs using Artificial Intelligence (AI) are gender equal. Empirical studies have already shown that the predominantly male computer science workforce has a reinforcing and perpetuating effect on the gender gap in AI development, which in turn increases the likelihood of producing algorithms with unconscious biases built into the data processing systems, which can further exacerbate the results of data treated through algorithms and other AI tools. Another worry is that these stereotypes have reached such a degree of permeation in society that female tech founders’ capacity to attract investors and customers is impaired. And that is just the beginning of the list of worries.

While most of the attention has been devoted of late to correcting gender (and other) biases found in AI applications, the deeper and more important challenge may be ensuring diversity and representativity in the creation of the AI platforms. Attaining diversity in the AI industry would certainly contribute to the goal of reducing gender bias by, for instance, making sure that women and members of other underrepresented groups have a voice in the design and systemic architecture of the technology being used to make more and more decisions that affect everyone’s day-to-day life. This goal would be served by consciously working towards recruiting and promoting more female and non-binary talent; by making it easier for female founders and entrepreneurs to access finance and support; by maximizing the diversity of images of AI products and robots in the media that undermine gender stereotypes, among other measures. Since negative gender stereotypes in advertising influence girls’ perceptions of STEM (Bond, 2016) and their subsequent career decisions (Cheryan, Plaut, Handron et al, 2013), they ultimately play a role in reinforcing and perpetuating the gender gap in AI. Given the increasing potential for real harm from algorithmic biases, now is the time to address the need for advertising and marketing strategies that represent the diversity (of gender and beyond) in our societies. In the end, the social fabric that unites society behind our common goals is nothing more than the sense of belonging and identification with society in general that is felt by the individuals that comprise it.

María Guadalupe Martínez Alles is an Assistant Professor of Law at IE University.