Some Black LinkedIn users complain of algorithmic bias, saying their posts are not appearing on the site or are being taken down repeatedly
Steven Melendez / Fast Company : Tweets: @kschwabable and @hypervisible . Thanks: @kschwabable Tweets: Katharine Schwab / @kschwabable : Our team's latest: A look at claims of censorship from Black LinkedIn users in the ad industry, which LinkedIn says are groundless. But there's no denying that the company is facing a perception problem that it's biased. Great reporting by @smelendez. https://www.fastcompany.com/ ... Hypervisible Dot Pdf / @hypervisible : LinkedIn doesn't deny that Black users' content is sometimes deprioritized, but rather than racism they claim it stems from “a mix of content-neutral technical bugs, the popularity among other users of particular posts, and the occasional moderation error.” 🙄 https://twitter.com/... Thanks: @kschwabable
Context & Ripple Effects
The complaints arrive a year into Black LinkedIn's emergence as a professional analog to Black Twitter, when pandemic-era activity and George Floyd coverage turned the network into a venue where moderation disagreements first surfaced. They also land against an awkward backdrop: LinkedIn's own CEO had already apologized months earlier after anonymous staffers echoed BLM skeptics at an internal town hall on racial injustice.
What makes the new reporting harder to dismiss is that LinkedIn does not deny deprioritization happens — it attributes it to bugs, popularity metrics, and moderation errors rather than racism. That distinction matters because [[a:967786|rivals like ZipRecruiter and CareerBuilder are already being pressed to strip gender and racial bias from their AI-powered matching]], and ranking systems on professional networks decide who gets seen, hired, and sponsored.
First-order effects
- Black users in the ad industry whose posts vanish or get repeatedly taken down lose visibility precisely where professional opportunity compounds — LinkedIn's 'groundless' denial leaves them without a working appeals path.
Second-order effects
- A perception-of-bias problem forces LinkedIn to treat moderation errors and popularity-metric quirks as audit-worthy engineering issues, the same class of fix job-search rivals are making to their matching algorithms.
Third-order effects
- If opaque ranking and takedown systems keep producing disparate outcomes on professional networks, expect pressure toward explainability requirements — content increasingly functions as inference input, and users will demand to know what their posts are inferred to mean.
The trend: Professional platforms are moving from treating algorithmic disparities as technical accidents to facing accountability demands over how ranking and moderation systems treat underrepresented users.