As AI-generated content floods the web, the only content that gets cited is content with an original perspective, first-hand experience, or a contrarian point of view. We also explain why user posts on Reddit and similar platforms have become a goldmine for AI citations.
What net information value means: It is the measurable value a piece of content adds to a reader’s understanding beyond the information that already exists on the same topic. If your article did not exist, what would readers lose? If the answer is “nothing,” the article’s net information value is zero.
With AI tools now everywhere, anyone can mass-produce articles with tidy headings and a clear conclusion. The web is flooded with articles that tick every structural box (headers, subheaders, a definition section, an FAQ, a summary) while containing almost no original thinking.
Writing “what is” and “how to” content purely for its informational value no longer works as a strategy. The market is saturated. Content that cannot be told apart from competitors’ gets chosen by neither AI systems nor people.
In response to AI, many brands and writers have turned defensive and started hiding their proprietary information. But with everyone playing it safe, all content has begun to look the same. When everyone settles for the safe summary, content that takes a clear stand and says something people could disagree with stands out sharply.
This is also why thought leadership from executives and founders is booming. Content in which a CEO writes about the challenges facing their industry carries an original perspective that cannot be generated from the existing information in an LLM’s training data. It is a textbook example of content with high net information value.
LLMs actively draw on UGC platforms such as Reddit, Quora, and Stack Overflow as citation sources. Why? Because these platforms host discussions in which real people share real opinions. They are messy, sometimes wrong, and often biased, and that is exactly why they carry far more informational value than a polished corporate blog that says nothing.
| Dimension | Low net value | High net value |
|---|---|---|
| Perspective | Both-sides hedging: “Experts are divided on this” | “Here is what we have seen work, and here is why the conventional wisdom is incomplete” |
| Sources | The same statistics and studies everyone else cites | Proprietary data, in-house research, or first-hand information only insiders have |
| Author expertise | A generalist who did two hours of research, or an AI | A practitioner who has actually done the work and can explain why the conventional wisdom is wrong |
| Replicability by AI | Replicable: the same prompt can generate an equivalent article | Not replicable: it rests on experience and perspective that do not exist in the training data |
AI Overviews, ChatGPT citations, and Perplexity answers: each of these systems runs its own “information value calculation” to choose content. They are looking not for the most average answer, but for the most specific and trustworthy one.
It may sound paradoxical, but much of the advice on “AI-optimized content” is counterproductive. Choose genuine specificity over templated structure, genuine opinions over vague phrasing, and depth in the areas you truly know over breadth that tries to cover everything. That is how content actually earns AI citations.
A question for you: When was the last time your company published something that made readers think, “I hadn’t thought of it that way”? If you struggle to answer, that is your opportunity.
FREE TRIAL
Check Your AISEO Score for Free
Just enter your URL to see how visible your brand is in AI search right now.
Get a Free Diagnosis →
![How Social Media and Content Formats Shape AI Citations [Research Report]: Inside a 3.8x Rise in Monthly Citations](/img/column/en/social-ai-citation.svg)
