LinkedIn search engine favours men's names

Occurred: August 2016

A newspaper investigation found that LinkedIn's search engine algorithm was suggesting male names when people were searching for female users, resulting in accusations of stereotyping and bias. 

According to an August 2016 Seattle Times report, a search of popular female first names, such as Stephanie and Andrea, were shown the result 'did you mean Stephen' or 'did you mean Andrew'. 

According to LinkedIn, its 'did you mean' results were produced by an algorithm designed to suggest names with similar spellings based on how frequently names have shown up in past queries. 

The company subsequently rolled out an update to the algorithm that enabled it to explicitly recognise popular names, so that the algorithm doesn’t try to correct them.

System 🤖

Operator: Microsoft/LinkedIn
Developer: Microsoft/LinkedIn

Country: USA

Sector: Business/professional services

Purpose: Augment search results

Technology: Search engine algorithm; NLP/text analysis; Machine learning
Issue: Bias/discrimination - gender

Transparency: Governance; Black box

Page info
Type: Incident
Published: March 2023