AI and LinkedIn ·

LinkedIn is starting to limit generic AI-created content

LinkedIn is adjusting its feed to reduce the visibility of low value content generated or amplified with artificial intelligence. The news has caused some confusion: it does not mean LinkedIn is going to penalise any use of AI, or that writing with the help of tools is a problem in itself.

The change is aimed at something else: posts, comments and videos that look mass manufactured, that repeat formulas, that bring no real experience and that degrade the quality of professional conversation. In other words, LinkedIn does not appear to be going after AI. It is going after empty content.

By Sheena de PunkVoice · Edited by Mario Pérez

Flat illustration of a person writing by hand at a desk on one side, while on the other a tall tower of identical grey sheets rises out of a pot, with a bookshelf, a plant and a terracotta mug in the room, evoking keeping your own content against discarded generic AI material.

The problem is not using AI, it is publishing without judgement

AI has made creating content easier. It has also made a lot of content start looking too alike.

Posts with identical structures. Sentences that sound good but say little. Generic advice with no context. Comments that repeat what the original post already said. Videos designed to capture attention but with no clear professional value.

For a while, this kind of content could work because it helped maintain a presence and a frequency. But if LinkedIn starts detecting those signals better, publishing more will stop being an advantage when what you publish adds nothing recognisable.

The important question changes. It is no longer enough to ask "am I publishing frequently?". The question should be "am I adding something that makes sense from my experience, my judgement and my professional position?".

Which kinds of content may lose visibility

LinkedIn has not published a complete technical list of factors, but the kind of content that appears to be under review follows fairly clear patterns.

  • Generic content. Posts that hundreds of different profiles could sign with barely any changes.
  • Repetitive content. Recycled ideas, predictable formulas and structures that have become too recognisable.
  • Content with no real experience. Advice that does not come from something lived, a concrete observation, a professional case or a reasoned position.
  • Automated comments. Replies that look correct but add no conversation, nuance or reading of their own.
  • Attention bait. Videos or posts designed to hold attention with no clear professional value.

The linguistic marks that give away a post made with ChatGPT

What changes for people building a personal brand

This move directly affects personal branding on LinkedIn. Until now, many strategies have leaned too heavily on frequency: publishing several days a week, commenting a lot, reusing formats and keeping constant activity. Frequency still matters, but it can no longer compensate for a lack of substance.

A solid personal brand needs something more:

  • A recognisable voice.
  • Concrete experience.
  • A point of view.
  • Recurring topics.
  • Coherence between who the person is, what they know and what they publish.
  • The ability to contribute a reading that is not interchangeable.

The content that best withstands the change

It is not necessarily the most elaborate. It is the content conveying that a person with judgement is behind it. Sometimes that shows in a well reasoned opinion. Other times in a concrete professional story. It can also appear in a simple but original observation. What matters is that the content does not look like one more piece off a production line.

AI can help, but it cannot replace the voice

Using AI to create content is not a problem. It can be useful for organising ideas, summarising notes, improving clarity, reviewing tone or turning an initial reflection into a cleaner draft.

The problem appears when AI takes all the important decisions:

  • What gets said.
  • The position it is said from.
  • What experience supports it.
  • What nuance makes it yours.
  • What stance sits behind it.

Why identity matters more and more

When those decisions disappear, the content can sound correct, even professional, but it loses identity. And on LinkedIn, identity matters more and more.

The reader is not only after well written sentences. They are after signals of experience, judgement and trust. They want to understand how a person thinks, not just read a polished version of ideas they have already seen many times.

LinkedIn prompts: when they help and when they dilute you

The new advantage will be publishing with more intention

This change can be positive for people using LinkedIn seriously. If the platform reduces the reach of generic content, posts that genuinely add something will gain weight: an experience, a reflection, an interpretation or a way of looking at a topic that is your own.

That means doing better work before publishing:

  • Choosing fewer topics, but clearer ones.
  • Avoiding overused formulas.
  • Checking whether the content fits your professional track record.
  • Asking whether anybody could sign the text.
  • Adding context, examples and real nuance.
  • Using AI as support, not as a substitute for thinking.

Accumulating posts does not build a brand

A personal brand is not built by accumulating posts. It is built by consistently repeating a way of thinking, working and adding value. The more the algorithm levels generic content, the more sustaining a recognisable voice over months pays off.

What to do now

LinkedIn's change does not demand abandoning AI. It demands using it with more judgement. Before publishing, it is worth reviewing a few questions:

  • Is this content connected to my real experience?
  • Does it bring a concrete idea or only repeat a formula?
  • Does it have a recognisable voice?
  • Does it fit my main topics?
  • Does it help my audience or only fill the calendar?
  • Could anybody else sign it without it showing?

If the last answer is yes, more work is needed

When anybody could sign a text, it probably needs a real example, a clearer opinion, a concrete reference, a lived experience or simply less generic writing. You do not always have to rewrite the whole post. Often changing the first sentence and adding a detail only you could contribute is enough.

Less volume, more voice

LinkedIn appears to be sending a clear signal: professional content has to add real value. AI will keep being part of the creation process. But using it indiscriminately can make many personal brands lose exactly what they should be building: a voice of their own.

LinkedIn's future will not be in publishing faster, it will be in publishing with more intention.

For people who want to use AI without falling into generic texts, tools focused on voice, editorial judgement and consistency can be especially useful. That is the space PunkVoice works in: helping build LinkedIn content without losing what makes a person recognisable. Because in a feed full of correct content, the difference will be sounding less interchangeable.

A comparison of personal branding tools for LinkedIn

Frequently asked questions

Is LinkedIn banning AI made content?

No. LinkedIn does not penalise using AI as support for writing or organising ideas. What it reduces in reach is generic, repetitive content with no real experience, regardless of whether an AI or a person wrote it. AI used with judgement remains legitimate.

How do I know whether my content counts as generic?

Ask yourself a simple question: could anybody else in my industry sign this post without it showing? If the answer is yes, it probably lacks concrete experience, an opinion of your own or a detail only you could contribute.

Which kinds of post may lose visibility?

Posts with recognisably identical structures, generic advice with no context, automated comments or ones that repeat the original post, videos designed only to hold attention with no clear professional value, and long series of posts with predictable formulas.

How can I keep using ChatGPT or Claude without it showing?

Use them before writing, not during or after: ask them to question your idea, spot clichés, suggest opposing angles. Then write the post yourself with those references. Never copy whole generated sentences, rewrite them in your vocabulary. And review the result out loud before publishing.

My reach has dropped lately. Is it because of this change?

It may be, but it could also be seasonal changes, adjustments in your own content or normal algorithm variation. Before assuming a penalty, audit three months of posts: if you recognise generic patterns or overused formulas, try publishing two posts with your voice clearly recognisable and observe the effect three or four weeks later.

How can PunkVoice help with this change?

Sheena, PunkVoice's editorial AI, is designed so AI does not decide for you. She reads your draft, spots clichés, questions ideas and reinforces your consistency, but she does not write the post in your place. It is exactly the kind of AI use LinkedIn's new feed approach rewards: an amplifier of judgement, not a substitute.