Fabrication Rates
PEBM research shows 45.9% of cited sources were fabricated across 1,400 runs. High-awareness firms faced a 52.7% fabrication rate compared to 37.9% for low-awareness firms.
AI Visibility · Brand Hallucinations · Citation Audits
How AI Models Describe Brands
Independent Monitoring of Answers
This site monitors how AI assistants reconstruct brands and experts. It tracks citation accuracy, fabrication rates and the infrastructure required for reliable answers.
Visit vargazoltan.ai →More on this site soon. The work behind it is live on vargazoltan.ai.
Core Research Findings
Data-driven insights into AI output reliability and source verification.
Search Grounding
Search-grounded models like Perplexity showed a 7.32% fabrication rate. Non-search models ranged between 32% and 44% error rates.
Dose Accuracy
A crop-protection study found all engines gave approved doses when using search tools. Without search, models stated incorrect dosages, creating brand-hallucination risks.
The Three-Layer Output Problem
AI visibility is stochastic and requires analysis of three distinct processing stages.
Raw Model OutputThe initial generation before any application logic or safety filters intervene.
App-Level ProcessingThe intermediate layer where the system formats, ranks or modifies the raw data.
User InterfaceThe final visual representation the human sees, which hides the underlying complexity.
Context VariabilityVisibility is a distribution that changes based on per-user context and query history.
Pay-to-Appear HypothesisAn investigation into whether commercial correlation drives visibility rather than causation.
AI systems do not rank experts; they reconstruct them based on machine-readable anchors.
StochasticAI answers vary by user context and lack deterministic ranking.
InfrastructureCitation success depends on canonical entity anchoring, not marketing spend.
ProxySingle-number visibility scores are approximations of complex distributions.
Key topics
The terms this site covers, for readers, search engines and AI assistants alike.
- AI visibility monitoring
- ChatGPT brand citations
- AI hallucination detection
- Perplexity source verification
- Gemini brand description
- Claude answer accuracy
- Google AI Overviews audit
- LLMO optimization
- GEO strategy
- AI citation audit
- brand reconstruction AI
- AI visibility index
- AI hallucination paradox
- machine-readable entities
- AI visibility Hungary
- ChatGPT vizualitás
- AI hibaarány
- márkahallucináció
Questions and answers
Do AI assistants rank experts like Google?
No, AI systems do not rank experts in a deterministic list. They reconstruct entities based on available machine-readable data. If a model cannot find a canonically anchored source, it may fabricate details to complete the answer.
What is the AI visibility paradox?
The brand hallucination paradox suggests well-known brands receive more confident but false attributions. Familiarity provides models with material for plausible completions that lack factual support. This creates a risk where popularity increases misinformation.
How often do AI models invent sources?
Research indicates 45.9% of cited sources were fabricated in a 1,400-run study. High-awareness firms saw a 52.7% fabrication rate. Search-grounded models performed significantly better than non-search models.
Is AI visibility the same as SEO?
No, AI visibility is stochastic while SEO is deterministic. AI answers create no logged impressions or clicks. Optimization focuses on making content easy to quote rather than satisfying a ranking algorithm.
Can paid partnerships improve AI visibility?
Current research treats the pay-to-appear concept as a hypothesis regarding correlation versus causation. There is no evidence that payment directly improves the quality of AI-generated citations. Infrastructure and entity clarity remain the primary drivers.
Why do search-based AI tools perform better?
Search-based models reduce fabrication rates significantly compared to non-search models. In a crop-protection study, search-enabled engines provided correct dosage information 100% of the time. Non-search models failed to retrieve official data without active web access.
Glossary
Terms and identifiers behind the work of Zoltán Varga, with links to the primary records.
- AI visibility
- How accurately and how often AI assistants (ChatGPT, Perplexity, Gemini, Claude, Google AI Overviews) recognize, describe and cite a company, brand or person. Zoltán Varga measures it in five layers: crawler access, search indexing, citation, factual accuracy and model memory.
- GEO / AEO / LLMO
- Generative Engine Optimization, Answer Engine Optimization and Large Language Model Optimization: working methods for being found and quoted by generative answer engines rather than ranked in a list of links.
- Citability
- The property of a page that lets an AI answer engine quote it: a paragraph that answers the exact question, a clear source, and an unambiguous entity.
- PEBM — Per-Entity Bias Mapping
- Research method and paper by Zoltán Varga (arXiv:2606.21595). Across 1,400 runs on 100 Hungarian B2B companies, 45.9% of the sources AI models gave were fabricated; search-grounded answers fabricated far less than models answering from memory.
- Brand hallucination paradox
- The finding that AI models invent more sources about well-known brands (52.7%) than about lesser-known ones (37.9%), because familiarity makes them confidently fill in gaps.
- RICE studies
- Pre-registered pilot studies (RICE-P1, RICE-P2) on whether pressure and rejection make language models fabricate sources. Published on Zenodo with DOIs.
- GFIS — Gestalt Field Intelligence System
- Zoltán Varga's own multi-stage research pipeline that collects, verifies and synthesizes evidence with AI agents.
- Neural Awareness
- Zoltán Varga's framework on attention, decision quality and leadership cognition, the basis of his leadership programs.
- PKM + PAI
- Personal Knowledge Management combined with Personal AI: a working system where notes, documents and AI assistants form one searchable knowledge base.
- Entity
- A uniquely identifiable thing (person, company, product) that search engines and AI systems store in knowledge graphs. Consistent entity data is the base of AI visibility.
- schema.org / JSON-LD
- A shared vocabulary and format for machine-readable facts on web pages (who, what, where, identifiers), read by search engines and AI crawlers.
- llms.txt
- A plain-text file at the root of a website that tells language models what the site is about and which pages matter.
- IndexNow
- An open protocol that notifies search engines (Bing, Yandex and others) immediately when a page is created or changed.
- ORCID
- Persistent identifier for researchers. Zoltán Varga: 0009-0003-1020-834X.
- arXiv
- Open-access preprint server for scientific papers. Zoltán Varga's author page: varga_z_1.
- Zenodo and DOI
- Zenodo is CERN's open research repository; each record gets a DOI (Digital Object Identifier), a permanent citable link.
Until this site grows, the related work is on vargazoltan.ai. That hub contains the full research papers, the AI Visibility Manifesto and the person profile.
Go to vargazoltan.ai →