The Real Problem With AI Content

You can generate content today without a human touching it, and it can rank surprisingly well — in AI search, even in Google. With a process and skills, pure AI content can get you decent results.

There's a catch though: that advantage shrinks as more people do the same thing.

The problem

Let's say you want to publish a piece on content marketing ideas for B2B SaaS.

You type the prompt, take what comes out and publish.

Countless teams can now do exactly the same thing and publish roughly the same article.

AI writes like one person

In a 2025 independent audit, Pangram was the best-performing AI detector tested, with essentially zero false positives and false negatives on medium and long passages.

But the accuracy isn't the interesting part. What's interesting is why detection works at all.

For every human document in its training set, Pangram generates an AI mirror — matched on topic, tone, and style — so the classifier can only learn the tells of LLM writing itself. It works, which proves the tells exist. AI has one recognizable style.

Why the style converges

Pangram's Joe Stech argues that pre-training gives a model the full range of human writing, then alignment crushes it toward an Annotator Consensus Dialect.

Idiosyncratic writing gets 5/5 from one human rater and 2/5 from another — a hedged, symmetrical answer gets a safe 4/5 from everyone.

Optimization collapses the variance into one smoothed voice. Academic research backs it up: RLHF causes mode collapse that persists across prompts, and stylistic diversity is already declining on Reddit and in scientific writing. So when you and I both ask an LLM to write, we publish the same person's writing.

What that means for ranking

AI assistants need a reason to cite you.

A model has no reason to cite a page it could generate itself.

The strongest reason to cite you is that you provide evidence the answer needs to attribute: original data, firsthand findings, or verifiable primary sources. Models can synthesize common knowledge. They still need sources for claims they can't independently verify.

Google measures originality.

The 2024 Content Warehouse leak contains an internal OriginalContentScore. Google doesn't treat uniqueness as a nice-to-have. It measures and ranks for it.

Google says it has ranking systems specifically designed to surface original content prominently, ahead of pages that merely cite it.

Its quality guidance asks whether a page contributes original information, research, or analysis and whether it provides substantial value beyond competing results. Google doesn't merely want another competent answer. It wants a reason to choose yours.

What can we do about it

There are essentially two problems to solve:

1. The writing isn't unique. By default, an LLM writes like the one writer the entire planet hired.
2. The research isn't unique. Ask an LLM to research a topic and it runs roughly the same searches, reads the same sources, and returns the same summary it gives everyone who asks.

Fix the writing

Dictate, don't prompt.

Talk through what you think, then let the model clean it up. This article started as a long voice note.

Write your voice down.

Build a brand-voice doc with real examples from your own writing and always check against it.

Fix the research

Map the field first.

Read what already ranks and name, specifically, what your piece adds.

Publish your own numbers.

Original data is the one input a model can't generate. The GEO study measured up to 40% more visibility in generative engines from adding statistics.

Skip the default research path.

Don't ask one model for "research on X" — that is what everyone else is doing.

Use different search systems such as Exa and Parallel Web to take different retrieval paths through the web instead of repeatedly following the same search results.

Send agents at the topic from several specific angles, force them into primary sources, and make them report what the top-ranking pieces don't say.

Epilogue

AI didn't kill content. It repriced it: sameness now costs nothing and is worth nothing. Uniqueness is the only thing left.

Notes

• Pangram. AI content detector, self-reported ~1-in-10,000 false-positive rate – pangram.com
• Jabarian, Brian, and Alex Imas. "Artificial Writing and Automated Detection." Becker Friedman Institute Working Paper, 2025 – bfi.uchicago.edu
• Russell, Jenna, Marzena Karpinska, and Mohit Iyyer. "People Who Frequently Use ChatGPT for Writing Tasks Are Accurate and Robust Detectors of AI-Generated Text." ACL 2025 – arxiv.org/abs/2501.15654
• Pangram Labs. "How AI Detection Works" – pangram.com/research
• Pangram Labs. "Why Perplexity and Burstiness Fail to Detect AI" – pangram.com/blog
• Stech, Joe. "The Information Theory Behind Why AI Writing Sucks." Pangram Labs, 2026 – pangram.com/blog
• "Understanding the Effects of RLHF on LLM Generalisation and Diversity." arXiv 2310.06452 – arxiv.org/abs/2310.06452
• "The Homogenizing Effect of Large Language Models on Human Expression and Thought." arXiv 2508.01491 – arxiv.org/html/2508.01491v1
• "Homogenizing Effect of Large Language Models on Creative Diversity." ScienceDirect, 2025 – sciencedirect.com
• Google. "A Guide to Google Search Ranking Systems" – developers.google.com
• Google. "Creating Helpful, Reliable, People-First Content" – developers.google.com
• King, Mike. "Inside the Google Algorithm Leak." iPullRank, 2024 – ipullrank.com
• Ahrefs. "Why ChatGPT Cites the Pages It Cites" – ahrefs.com/blog
• Aggarwal, Pranjal, et al. "GEO: Generative Engine Optimization." arXiv 2311.09735 – arxiv.org/abs/2311.09735
• "Google's March 2026 Core Update Shifted Visibility Away From Aggregators." Search Engine Journal – searchenginejournal.com
• Walter Writes AI – walterwrites.ai
• Exa – exa.ai
• Parallel Web – parallel.ai