Addressing gender considerations in AI-driven ads #IWD2025

AI is increasingly integral to paid media, automating processes and boosting efficiency. However, the reliance on historical data can inadvertently perpetuate existing societal biases, particularly concerning gender. It’s important to understand how these biases manifest and how we can mitigate them.

Observed trends in ad platforms

Meta & LInkedin

• Research from Cornell University indicates that Meta’s algorithms may disproportionately display high-paying job advertisements to male users, potentially limiting female users’ awareness of such opportunities.
• Similarly, LinkedIn’s algorithms have been observed to potentially filter out female users from advertisements in specific industries, such as tech and engineering, based on historical application data.
• Northeastern University’s findings suggest that Meta’s ad delivery can be influenced by image content, with ads featuring young women sometimes being disproportionately delivered to older male users.

Google

A study by Carnegie Mellon University revealed a potential bias in Google’s advertising system, where simulated male users were more likely to be shown advertisements for high-paying executive positions compared to simulated female users.

Practical steps for mitigating potential bias

• Data input practices: Utilise diverse and gender-balanced datasets when creating custom audience lists, ensuring equitable representation.
• Audience review: Regularly audit lookalike audiences to confirm gender inclusivity.
• Visual content: Employ diverse and inclusive imagery that challenges traditional gender stereotypes, particularly in industries where representation has historically been skewed.
• Ethical AI: Advocate for the development and implementation of ethical AI practices within advertising platforms, encouraging the use of inclusive datasets.

Key takeaway

While AI offers significant advantages, it’s crucial to acknowledge the potential for bias. By proactively addressing these concerns through data practices, visual representation, and platform engagement, we can strive to create a more equitable and inclusive advertising environment.

#IWD2025