Habr AI explained why social media users react so strongly to AI-written texts

Negativity towards AI-generated content on social media is not only explained by text quality. The column author explores a more unpleasant reason: generative posts destroy the already shaky illusion that the endless feed carries some personal meaning.

Why this angers

The phrase "another GPT post" increasingly sounds not as a technical comment, but as an instant verdict. A user sees not simply text that an algorithm could have written, but another unit of content assembled for reach, comments and feed retention. Even if the post is formally neat, the mere suspicion of machine origin changes its status: instead of a thought or experience, a person finds themselves facing an impersonal tool for algorithms.

According to the author's observations, this is particularly noticeable in Threads, where the assumption "AI wrote this" often sounds without direct insult, but still carries a negative meaning. People seem to instantly decode: if the text is machine-assembled, then the author either couldn't be bothered, or never intended to share something important. What bothers them is not only stylistic smoothness, but the feeling of substitution — as if instead of a conversation partner, a template was quietly put into the conversation, one that should simply hold attention for a couple more seconds.

Need for author's filter

At the same time, generation itself is not declared evil. The author gives examples of AI music and AI video that work as full-fledged statements: behind them you feel intent, selection and taste. If a person uses a model as a tool, and then decides what is worthy of publication, there are fewer complaints.

The problem begins where generation looks like a stream of raw material that no one cut off, reassembled or passed through their own position. This logic works outside the AI world too. Content overload annoys even from live authors if they publish everything in a row and shift the burden of choosing what's good onto their audience.

For a reader, what matters is not the biological status of the creator, but the presence of an editorial filter. They want to understand why exactly this text, track or video appeared in their feed, why it was chosen, and what is worth their attention in it.

• Does the publication have an original thought or emotion
• Did the author select the best result, not the first generated variant
• Is it visible that the material was worked on, not dumped in one click
• Is the post trying to communicate something, or just hook the algorithms

Three legs of the feed

The key thesis of the article is that the problem runs deeper than AI content itself. Social media has long rested on three mechanisms: variable reward, habit and attention to signals from other people. A user scrolls through the feed hoping to catch something valuable, as if panning sand searching for gold. At one time this was at least partly justified by the presence of friends, colleagues and acquaintances, but over time they were displaced by influencers, brands and endless content factories.

"Consumption of content served by social networks makes no sense at all."

When generative text massively enters this system, the last support breaks — the feeling that the signal after all comes from a human. If the feed is already filled with other people's faces, with whom we have no real connection, then AI posts finish off the remaining illusion of sociality. Hence the sharp reaction: this is not simply a dislike of new technology, but an unpleasant realization that what we have before us is no longer a conversation, but an endless conveyor of stimuli.

What this means

For authors, brands and media, the conclusion is unpleasant but useful: simply "well-written" AI text is no longer enough. Audiences are increasingly faster at recognizing content made without authorial participation, and treat it as noise. This means value shifts not in generation itself, but in selection, editing, position and a clear answer to the question of why this material appeared in the feed at all and what it gives to a living reader.