AI in advertising: tackling harmful gender biases

As Paid Media Specialists, we frequently utilise AI to enhance and optimise advertising campaigns. Ad platforms have made great progress in using machine learning to automate targeting, bidding and even creative development. While these technologies offer efficiency and effectiveness, they are not without their drawbacks. One major concern is gender bias. AI is trained on historical data, which often contains societal biases. These biases, grounded in traditional gender roles and stereotypes, can influence how ads are targeted and the content they display. A cycle is created where the AI perpetuates these outdated ideas. It needs to stop.

Ad platforms reinforcing harmful gender bias

Meta Ads

A study by Cornell University found that Meta's algorithm showed job ads for senior and higher-paying roles to men more often than women, even when they had the same qualifications. This kind of targeting further entrenches general inequality by making it less likely for women to be aware of opportunities. Even more troubling, LinkedIn's algorithm was found to filter out women when advertisers targeting specific industries, for example, women in the UK might be less likely to see ads for tech or engineering jobs, simply because the system was biased towards showing these ads to men, based on historic data about who had previously applied for such roles.

Similarly, research from Northeastern University found that Meta's ad delivery algorithm discriminated based on gender, race and age, influenced by the images used in ads. Ads featuring women were generally shown more often to women, however ads featuring young women were disproportionately delivered to older men, a phenomenon the researchers dubbed "Creepy Old Man Effect". This illustrates how algorithms not only reflect but can amplify societal biases.

Google Ads

A study by The Carnegie Mellon University investigated gender bias in Google Ads. The research found that Google's advertising system was biased in favour of showing high-paying job ads to men over women. Specifically, the study showed that simulated male users were more likely to see ads for high-paying executive positions, while simulated female users were shown ads for lower paying roles. The research highlighted the potential for online ad algorithms to perpetuate gender inequalities in the workplace, especially in high-paying sectors.

How can we tackle gender bias from AI in ad platforms?

As AI continues to play a larger role in advertising, it is crucial we address the risk of gender bias in our campaigns to prevent reinforcing biased AI models. Below are some actionable steps advertisers can take to minimise bias and enhance the inclusivity of AI-driven advertising.

1. Feed Platforms With Accurate Data:

By feeding platforms with diverse, gender-inclusive data, advertisers can help combat gender bias within advertising platforms. For example, custom lists that include equal representation of both genders will help algorithms learn to target both men and women effectively and fairly. The more inclusive your data input, the better AI will perform, and the less likely it is to fall back on outdated gender biases.

2. Examine Audience Settings

Lookalike audiences are based on data you provide, so it's important to check that both male and female data points are included.

3. Use Diverse Imagery

Imagery plays a crucial role in shaping perceptions. Make sure campaigns use diverse inclusive imagery that does not conform to outdated gender roles. Where creative is limited, stock imagery choices become especially important. Use images of women in leadership positions for tech and STEM roles and avoid reinforcing stereotypes of men in only executive positions. Diverse representation in stock imagery also speaks to a broader audience, enhancing brand credibility and appeal.

4. Advocate for Ethical AI in Advertising

It's essential for marketers to advocate for ethical AI practices within the platforms we use. Urge core platforms to use inclusive datasets and work towards reducing gender bias in advertising.

Summary

As AI-driven advertising continues to grow, it is essential for us to recognise the risks of reinforcing gender stereotypes. As female marketers and Paid Media Specialists, we hold the power to challenge and minimise these biases. By proactively auditing our campaigns, diversifying imagery and championing ethical AI practices, we can lead the way in creating a more inclusive advertising landscape that promotes quality and empowers women in the digital space.