LinkedIn’s feed has become a source of frustration for creators who rely on the platform for visibility, networking, and business growth. After the company said it had “more recently” implemented LLMs to help surface useful content, some heavy users began reporting sharp changes in engagement and impressions.
The most visible challenge came through #WearthePants, an experiment in which women tested whether LinkedIn’s new algorithm treated them differently when their profiles appeared male. The results were anecdotal, but they raised a larger question: when AI helps decide what gets seen, how much can users really know about why their posts rise or disappear?
What the #WearthePants experiment tested
One participant, a product strategist identified by TechCrunch as Michelle, changed her LinkedIn gender to male and changed her name to Michael. She said she joined #WearthePants because she suspected LinkedIn’s new algorithm was disadvantaging women.
Michelle had more than 10,000 followers and also ghostwrites posts for her husband, who has only around 2,000. According to her, both accounts often received around the same number of post impressions despite the difference in audience size. “The only significant variable was gender,” she said.
Other women reported similar experiences. Marilynn Joyner, a founder who had been posting consistently for two years, said visibility had declined in recent months. After changing her profile gender from female to male, she told TechCrunch, “my impressions jumped 238% within a day.”
The experiment began with Cindy Gallop and Jane Evans, who asked two men to post the same content as them. Gallop and Evans had more than 150,000 combined followers, while the two men had around 9,400 at the time. Gallop said her post reached only 801 people, while the man who posted the exact same content reached 10,408 people.
LinkedIn says gender is not a ranking signal
LinkedIn rejected the idea that its feed visibility is determined by demographic information. The company said its “algorithm and AI systems do not use demographic information such as age, race, or gender as a signal to determine the visibility of content, profile, or posts in the Feed.”
It also warned that a limited comparison of feed outcomes does not prove unfair treatment. LinkedIn said that “a side-by-side snapshot of your own feed updates that are not perfectly representative, or equal in reach, do not automatically imply unfair treatment or bias” within the Feed.
LinkedIn told TechCrunch that it tests millions of posts to connect users to opportunities. The company said demographic data is used only for testing, such as checking whether posts “from different creators compete on equal footing and that the scrolling experience, what you see in the feed, is consistent across audiences.”
The company also said its AI systems consider hundreds of signals, including a person’s profile, network, and activity. In a statement, LinkedIn said, “We run ongoing tests to understand what helps people find the most relevant, timely content for their careers.”
Why experts say the answer is complicated
Experts cited by TechCrunch did not treat the #WearthePants results as simple proof of explicit sexism. Brandeis Marshall, a data ethics consultant, described platforms as “an intricate symphony of algorithms that pull specific mathematical and social levers, simultaneously and constantly.”
Marshall said a profile photo and name are only one lever among many. She also pointed to user behavior and interactions as part of the system. “What we don’t know of is all the other levers that make this algorithm prioritize one person’s content over another. This is a more complicated problem than people assume,” she said.
That complexity matters because changing a profile during a viral trend could affect engagement in ways unrelated to gender. Some accounts were also posting after periods of inactivity, which could have influenced how the system responded.
Michelle herself changed more than her profile. She said that during the week she posted as “Michael,” she wrote in a more simplistic, direct style, similar to the posts she writes for her husband. She said impressions jumped 200% and engagements rose 27%.
Her conclusion was not that the system was “explicitly sexist.” Instead, she suggested that communication styles commonly associated with women may have become “a proxy for lower value.”
The AI ranking problem behind the complaints
The concern is not only whether LinkedIn directly uses gender. The larger issue is whether an AI-driven ranking system can reward patterns that reflect existing social bias, even without explicit demographic targeting.
Marshall said many platforms “innately have embedded a white, male, Western-centric viewpoint” because of who trained the models. TechCrunch also noted that researchers have found human biases such as sexism and racism in popular LLM models because they are trained on human-generated content, with humans often involved in post-training or reinforcement learning.
Sarah Dean, an assistant professor of computer science at Cornell, said platforms like LinkedIn often use full profiles and user behavior when deciding what to boost. That can include jobs on a profile and the kinds of content a person usually engages with. “Someone’s demographics can affect ‘both sides’ of the algorithm — what they see and who sees what they post,” Dean said.
That does not prove a single cause for the #WearthePants outcomes. But it does explain why users can experience the feed as unfair even when a company says demographic fields are not used for ranking.
Creators want clarity, but the feed is more competitive
The complaints are not limited to women. Shailvi Wakhlu, a data science consultant, said she has averaged at least one post a day for five years and used to receive thousands of impressions. Now, she said, she and her husband are lucky to see a few hundred.
One man told TechCrunch he saw about a 50% drop in engagement over the past few months. Another said his post impressions and reach increased more than 100% in a similar period because he writes on specific topics for specific audiences.
LinkedIn said its user base has grown, creating more competition in the feed. It said posting is up 15% year-over-year and comments are up 24% YOY. The company also said posts about professional insights and career lessons, industry news and analysis, and educational or informative content around work, business, and the economy are performing well.
For creators, that still leaves a practical problem. If reach rises or falls without a clear explanation, users cannot easily tell whether the cause is writing style, audience behavior, content category, profile signals, competition, or an AI update. Michelle summed up the frustration plainly: “I want transparency.”
That demand is difficult for platforms to satisfy. Content-ranking systems are closely guarded, and full transparency could make them easier to manipulate. But without more clarity, LinkedIn’s AI feed will likely keep facing the same question from creators: whether the system is rewarding value, or simply reinforcing patterns users cannot see.