OK, what’s going on with LinkedIn’s algo?
Meanwhile, a group of women conducted an experiment called #WearthePants to test the hypothesis that LinkedIn’s new algorithm was biased against women.
For months, some heavy LinkedIn users complained about seeing drops in engagement and impressions on the career-oriented social network.
One of the experiment participants, Michelle, a product strategist with over 10,000 followers, noticed that her posts were getting around the same number of impressions as her husband’s, despite his smaller following.
The Experiment
Michelle and other women changed their profile gender to male and reported significant increases in impressions and engagement.
Marilynn Joyner, a founder, saw her impressions jump 238% within a day after changing her profile gender to male.
Other women, including Megan Cornish and Rosie Taylor, reported similar results.
LinkedIn’s Response
LinkedIn denied any bias in its algorithm, stating that demographic information such as age, race, or gender is not used as a signal to determine the visibility of content.
However, social algorithm experts agree that implicit bias may be at work, and that the algorithm is influenced by various factors, including user interactions and content.
The Complexity of AI
Data ethics consultant Brandeis Marshall explained that the algorithm is an intricate symphony of mathematical and social levers that constantly interact and influence each other.
She noted that the changing of one’s profile photo and name is just one such lever, and that the algorithm is also influenced by how a user interacts with other content.
The Unknown Variables
Marshall said that the unknown variables are probably the reason why some women saw increased impressions after changing their profile gender to male.
She suggested that partaking in a viral trend, tone and writing style, and other factors may also play a part in the algorithm’s decision-making process.
Michelle concluded that the system was not explicitly sexist, but seemed to deem communication styles commonly associated with women as a proxy for lower value.
Source: Link
