> ## Documentation Index
> Fetch the complete documentation index at: https://docs.peec.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Wikipedia in AI search — influences models on two levels

> Peec's 30M-source analysis (Mar 2026) — Wikipedia is the #4 most-cited domain and shapes AI twice: live retrieval plus core training data, so a pre-cutoff page means the model already knows you.

Wikipedia influences AI answers on two distinct levels, per Peec's 30-million-source analysis ([full study](https://peec.ai/blog/top-domains-cited-by-ai-search-analysis-based-on-30m-sources), Mar 2026), where it ranks as the #4 most-cited domain overall — strong on ChatGPT (its #1 source) and Perplexity, absent from Google's top 5.

* **Live retrieval:** AI engines actively pull from Wikipedia when generating answers.
* **Training data:** Wikipedia is one of the core sources models learn from before release. A brand with a Wikipedia page before the model's knowledge cutoff is not just a citable source — it is something the model already understands.

The catch: Wikipedia has strict notability guidelines, and submissions that fail them get rejected or deleted. Not a quick win, but worth pursuing for brands that qualify.

How training vs retrieval access maps to different crawler bots (GPTBot vs OAI-SearchBot vs ChatGPT-User) is covered in [AI bot types](/research/ai-bot-types). Per-engine source rankings: [Most-cited domains by engine](/research/most-cited-domains-by-engine).

**Source:** [Top domains cited by AI search: Analysis based on 30M sources](https://peec.ai/blog/top-domains-cited-by-ai-search-analysis-based-on-30m-sources), Peec (Tomek Rudzki), Mar 2026.