The Truth About AI in Social Media

The way we use AI in our social production is simple: it's a thinking partner, not a content machine. When we're deep in a content sprint and we need to visualise our ideas en masse, AI helps us unblock. It's useful for brainstorming caption angles, sense-checking hooks, or exploring a topic from a different direction when we've been staring at a brief for too long.

What it doesn't do is write our posts. Because the moment it does, you lose the thing that actually makes content perform: a specific, human perspective. The opinion, the anecdote, the slightly unexpected take. That's not something you can prompt your way into. It has to come from a person who knows the brand, the audience and has something to say.

If you've noticed a dip in your LinkedIn reach recently, AI-generated content may well be the reason, and this isn't speculation. LinkedIn has rolled out its most significant algorithm overhaul to date, a 150-billion-parameter AI model called 360Brew, which has been steadily reshaping how content is distributed across the platform.

Here's the part that should make any brand sit up: 360Brew is specifically designed to detect and suppress generic, unedited AI output. We're talking about Natural Language Processing (NLP) classifiers that flag tell-tale signs. These are machine learning algorithms that analyse and assign predefined categories, tags or labels to text, which filter for spam, perform sentiment analysis and content moderation. For example, the "In today's fast-paced world" openers, the over-structured bullet lists, the templated frameworks that look the same account to account. Posts flagged in this way receive sharply reduced distribution right out of the gate. Some reports suggest organic reach has dropped by around 50% year-on-year, with AI-generated content taking the hardest hit.

Meanwhile, LinkedIn has also been expanding its verification filters to combat fake profiles and bots across feeds, comments, and job applications, all pointing to a platform that is actively investing in making authentic, human content the only content worth distributing.

The message is clear: the algorithm now rewards depth, expertise, and a point of view. Not volume. Not polish. Perspective.

So what does our process look like in practice? Here's an honest breakdown:

Where AI helps us:

• Supporting the generation of content angles to allow us to hone in on the most interesting, relevant one

• Rapid-fire caption variations to stress-test different tones

• Research prompts to pull together talking points before we write

• Identifying content gaps within a pillar strategy

Where a human takes over (every time):

• Choosing the angle that actually fits the brand's voice

• Writing the final copy, with the specific detail, texture, and tone that make it theirs

• Adding the real-world context, opinion, or insight that makes someone stop scrolling

• Deciding what not to post

The output that goes out into the world is always human-led. AI gets us to the starting blocks faster; it doesn't run the race.

AI isn't going anywhere, and we're not pretending otherwise. But the brands that will win on social in 2026 and beyond aren't the ones using it the most, they're the ones using it the smartest.

Use it to think faster. Use it to pressure-test ideas. Use it to draft things you'll then tear apart and rewrite. But never, ever let it replace the human brain behind your brand. Because that's the part no algorithm can replicate, and right now, it's also the part LinkedIn is actively rewarding.

Want a social strategy that's built around authentic, high-performing content? Let's talk.