LinkedIn's Algorithm May Be Filtering AI Posts, Cybernews Reports

LinkedIn is reportedly demoting AI-generated posts in its feed, according to a Cybernews analysis titled "Why AI-generated posts are failing LinkedIn's algorithms – and your brand." The piece frames the issue as a dual problem: weaker algorithmic reach and brand-level damage when content reads as machine-written.

What does the headline actually tell us?

Not much, on its own. The excerpt Reach Brief reviewed is a link plus byline, not the body text. We can't see Cybernews's sample size, the ranking signals tested, or the time window the analysis covered. That matters, because LinkedIn's own algorithm commentary has shifted across recent product updates, and third-party ranking research is rarely reproducible without the underlying data.

What does this change for social media teams?

If the pattern Cybernews describes holds, marketers using AI drafting tools at scale should expect softer organic impressions relative to human-edited posts. The deeper risk is cumulative. Accounts repeatedly publishing flagged content can pick up a quality penalty that depresses future reach across the whole page, not just one underperformer.

A few guardrails worth running before you overhaul the workflow:

• Compare impressions on AI-assisted posts against human-written posts from the same account over a 30-day window, with at least 20 posts per group for a meaningful sample.
• Watch LinkedIn Page Analytics for "sends" and engagement-rate trends; sends have historically correlated with feed ranking weight.
• Track personal-profile posts separately from company-page posts; LinkedIn has signaled different treatment for the two surfaces.

Is LinkedIn actually detecting AI text?

That's the open question. Detection could run through stylistic signals, post metadata, or aggregated user behavior like low dwell time. Cybernews's framing implies detection is happening, but the excerpt doesn't specify the channel, so treat the mechanism as unconfirmed.

LinkedIn's public guidance stays narrow. The platform allows AI-assisted posts under its User Agreement but prohibits misrepresenting AI content as human-authored. Most of LinkedIn's algorithm commentary targets spam signals — repetitive posting, engagement bait, low dwell — not AI specifically. So the honest reading: there is a real conversation happening about whether mass-produced AI posts are being quietly deprioritized, but the verifiable evidence, in what Reach Brief could review, sits in the headline rather than the body.

Should you stop using AI for LinkedIn drafts?

Not necessarily. The move with the lowest downside is a hybrid: human strategy, AI assistance for first drafts, manual edits, and a human voice check before publish. That workflow is most likely to clear whatever quality filters LinkedIn is running, whether those filters specifically target AI or not.

Watch for two things in the coming weeks. First, whether other industry outlets replicate the Cybernews finding with their own datasets. Second, whether LinkedIn publishes formal guidance specifically about AI-authored content ranking. Until one of those lands, treat the Cybernews report as a warning signal, not a confirmed rule.