Home / Technology / LinkedIn Gender Algorithm: Women Test Bias
LinkedIn Gender Algorithm: Women Test Bias
13 Dec
Summary
- Women changed LinkedIn profiles to male, observing engagement shifts.
- LinkedIn states its algorithm does not use gender for content visibility.
- Experiment suggests implicit bias may favor male communication styles.
A recent experiment, dubbed #WearthePants, saw several women switch their LinkedIn profiles to male. They reported a notable increase in post engagement and impressions, raising suspicions about algorithmic bias against female users on the platform. LinkedIn, however, maintains that its algorithm and AI systems do not utilize demographic information such as gender to influence content visibility.
Experts in social algorithms suggest that while explicit sexism might not be the cause, implicit biases embedded in AI training data could be at play. These biases may inadvertently favor communication styles stereotypically associated with men, leading to perceived disparities in content reach.
The experiment, initiated by entrepreneurs Cindy Gallop and Jane Evans, involved men posting identical content to women, who then observed significantly lower engagement. LinkedIn stated that its AI considers hundreds of signals, including user behavior and content history, to determine feed relevance, and that its tests aim for equal footing across creators.
