The problem

The problem with modern networking sites

From:
, founder
Published:
Updated:

The short answer: modern networking sites mostly make money from attention, so their feeds reward whatever grabs it. That pushes people towards performance (life lessons from a toddler's nap, "agree?" polls, humblebrags, AI-written think pieces) and away from the plain, useful work talk the sites were meant for. The fix is not a better algorithm for the same thing. It is a network with different rules.

What were professional networks for?

The original promise was simple. Put your work history somewhere sensible, find people in your field, hear about jobs, ask a question and get a straight answer from someone who knows. A digital version of the trade show, the industry newsletter and the colleague you could ring.

Some of that still works. Recruiters still search profiles. People still get hired through a message from someone they once worked with. But anyone who opens a professional feed today knows that is not what most of it is.

What goes wrong with the feed?

A feed has to decide what to show you first. Most decide by predicting what you will react to, because reactions keep you on the page, and time on the page is what advertisers and subscription upgrades pay for.

That single choice shapes everything people post. If a story about crying in a car park gets ten times the reach of a useful note on supplier lead times, people learn fast. They are not stupid. They are responding to incentives, the same way anyone would.

The result is a familiar set of genres:

  • The life lesson. A small domestic moment turned into a leadership principle.
  • The bait. "Comment YES and I'll send you the template." "Agree?"
  • The humblebrag. "So humbled to announce..." followed by something nobody asked about.
  • The holiday post. Your colleague's beach, framed as "what this trip taught me about resilience".
  • The machine-written essay. Long, smooth and empty.

None of these are crimes. Plenty are harmless. But together they crowd out the reason you opened the app.

How much of it is written by AI now?

A lot. Originality.ai, which sells AI detection, has tracked longer public LinkedIn posts for several years. Its study of posts up to October 2024 put the share that were likely AI-written at about 54%. Its larger July 2026 sample of 5,000 posts of 100 words or more classified 81.2% as likely AI. AI detectors are not perfect, and that study looks at a slice of public posts rather than every feed, but the direction is hard to argue with.

LinkedIn itself has said it is reducing the reach of generic, repetitive content that appears AI-generated, and has been rolling out a "seems like AI slop" feedback option. That is a sign of how big the problem has become.

Why do fake accounts matter?

A professional network only works if the people on it are real. LinkedIn's own transparency reports show the scale of the fight: in the first half of 2019 alone it removed more than 21 million fake accounts, and it has reported blocking tens of millions more in each half year since. Most are stopped at sign-up, which is good, but the ones that get through are the ones that send you the too-good-to-be-true job offer.

The root problem is that most networks let anyone sign up with any email address and type any job title. Checking is done afterwards, by machines and by members reporting things.

Why don't people just post better?

Because the feed punishes it. Research by Ovul Sezer, Francesca Gino and Michael Norton, published in the Journal of Personality and Social Psychology in 2018, found that humblebragging makes people less liked and seen as less sincere than plain bragging. People keep doing it anyway, partly because everyone around them does. When the norm on a platform is performance, opting out feels like disappearing.

So asking individuals to "just be authentic" misses the point. The rules of the room decide how people behave in it.

What would a fix actually look like?

We started Office Worker because we think the fix is structural, not a new feature. In practice that means a few rules:

  1. Real workers only. You sign up with a work email address. Personal email providers such as Gmail, Yahoo and Outlook are not accepted, so everyone on the network has a verified link to a real organisation. More on that in why we verify work emails.
  2. Work talk only. Every post is checked before it reaches a feed. Holiday photos, lunch and life lessons from your dog are not what the feed is for, so they are sent back before they reach anyone. Spam, harassment and misinformation are sent back too. You are told why.
  3. No bait economy. "Agree?" and "comment YES" posts are held for a person to look at before anyone sees them. Bait never earns extra reach.
  4. Reasons, not mysteries. When a post is stopped or held for review, you see the reason in plain words and can appeal. A person looks at appeals.

It is a narrower kind of network. That is the point. You can read what a work-only network looks like in more detail, or compare it with the other options in our guide to LinkedIn alternatives.

Is narrower actually better?

For some things, no. If you want to share your marathon time with everyone you have ever worked with, a general network is the right place. But if you want a feed you can open at 9am, read for five minutes and come away knowing something useful about your industry, fewer genres is a feature.

The best work conversations most of us have had were not performances. They were someone saying "here's what we tried, here's what broke, here's what we'd do differently." That is the bit worth building a network around.

Frequently asked questions

Why do professional networking sites feel so fake?
Most are funded by advertising and subscriptions, so they reward whatever keeps people scrolling and reacting. Posts that provoke a quick reaction, such as emotional stories, polls and bait, travel further than plain, useful work talk, so people learn to write for the feed rather than for colleagues.
Is it just LinkedIn?
LinkedIn is the biggest example because it has more than a billion members, but the pattern is the same anywhere a feed is ranked by engagement. Any network that pays out attention for reactions will fill up with content designed to get reactions.
What would a better professional network look like?
One where every member is verified as a real worker, posts are about work, the feed is not tuned for outrage or bait, and there is a clear, published reason when something is taken down.

Sources

  1. Originality.ai: LinkedIn AI content study
  2. LinkedIn community report
  3. Sezer, Gino and Norton (2018), Humblebragging: a distinct and ineffective self-presentation strategy