The problem with LinkedIn goes beyond AI

12 August 2026 - LinkedIn has spent the last few months reworking what gets rewarded in the feed. In May, the platform started reworking its algorithm to reward comments over likes and to penalise posts that people scroll past in under ten seconds. Not long after, its editorial team confirmed it's building classifiers to spot and suppress generic AI content specifically. As LinkedIn’s VP of Global Editorial, Laura Lorenzetti, confirmed in a company blog post: posts that repeat existing ideas without adding anything to them will be deprioritised, while also trying not to punish anyone who simply uses AI as part of how they write. A delicate balancing act considering the integration of AI into so many professions. Fundamentally, they're targeting content that's cosplaying insight, rather than actually sharing a worthwhile perspective.

It's tempting to read this as an AI story, but LinkedIn had a formula problem years before it had an AI problem. The rise-and-grind daily routine post, the "what shipping my first title taught me about resilience" post, the humblebrag success story, all inevitably ending with a tidy epiphany and a vacuous takeaway. None of that needed a language model to exist. It was already a genre, built by humans, long before anyone could generate it on demand. All AI did was provide a machine that could produce that cadence instantly and at volume. Slop isn't a new problem on any platform.

The fix people keep landing on - be more authentic - is only half right and slightly unhelpful as advice. Authenticity is important, yes, but the issue is that most people, the moment they open LinkedIn, switch into a separate voice that doesn't belong to them anywhere else. Nobody opens a conversation with a sweeping observation about the future of work. That register was invented for the platform, and it was always a bit hollow, and now large language models can produce it faster than a person can, which is exactly why it's started collapsing under its own weight.

A post only ever gets someone's attention. Business is built on whether you and your product or service deliver, so there's little reason to hide behind a safer version of itself. Confidence in your own perspective, expertise, and what you're selling is the thing a post is meant to signal, but that can be difficult to project when you're too close to your own expertise to hear how it lands from the outside

It is true that some elements of LinkedIn voice are best practice; it does work for driving engagement. Shorter paragraphs are easier to read; a good hook will encourage people to read on; calls to action will drive more people to wherever you want them to go; personal anecdotes are relatable. But the performance is more textural than that. There are ways to have all of the above and still sound like a human.

The fix for LinkedIn voice is simple: talk the way you'd talk at work rather than trying to out-human an AI. Still professional, and perhaps a bit cleaner than the version of yourself on Instagram or X, but just the version of you that shows up in a Slack message to a colleague, or in a client meeting, situations where you have an actual opinion to offer, couched in your expertise. That's a wider register than most people give themselves permission to use on the platform, and it happens to be exactly what the new algorithm is picking up on.

It's still an understandable tension that founders are facing. Nobody wants to say the wrong thing on a platform where developers, publishers and potential investors are all reading. This caution is exactly what's being penalised now, both by an algorithm that can smell a template from two sentences in and punishes inconsistency, and by an audience who are increasingly recognising performance for what it is.

None of this means abandoning judgement. Staying professional and having a voice aren't opposites, whatever the last decade of LinkedIn content might have suggested. The people whose posts are most interesting are the ones who don't code-switch the moment they open the app, and who are willing to say a specific, insightful thing rather than a safe, well-worn one.

It's hard to get right on your own, particularly when you're too close to the thing you're trying to talk about to hear how it sounds from the outside. That's where Raptor PR comes in, helping early-stage gaming leaders to help them build the industry credibility their expertise deserves and communicate it with confidence.

‍

By Sophia Hollis, Senior Account Executive