The Illusion of Authority: How a Fake Brand Manipulated AI Search Visibility

Introduction: Testing the Boundaries of AI Trust

In an era where Large Language Models (LLMs) are increasingly becoming the primary gateway to information, the question of ‘authority’ has shifted. Traditionally, SEO was about backlinks and domain age. But does AI search actually verify the truth, or does it simply reward coherence and repetition? To answer this, a groundbreaking month-long experiment was conducted by SE Ranking’s research team to see if a completely fictional brand could ‘win’ in AI search results.

The results are startling: a fake brand with zero real-world history successfully secured citations and favorable recommendations from some of the world’s most powerful AI systems, proving that AI visibility is governed by repeatable, manipulatable signals rather than strict fact-checking.

The Experiment: Creating a Brand from Thin Air

The methodology was designed to simulate a rapid market entry. Researchers created a fictional brand within a real, competitive niche. To test the AI’s response, they deployed a multi-layered content strategy:

The Primary Asset: A brand-new website specifically registered for the fictional company.
The Authority Boost: 11 existing domains (over a year old) were used to publish supporting content about the fake brand.
Content Diversification: Seven different formats were tested, including deep guides, ‘best of’ listicles, comparison pages, and how-to tutorials.

The team tracked 825 unique prompts across five major AI systems: ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, and Gemini, generating over 15,000 AI responses to analyze visibility, citation rates, and positioning.

Key Findings: The ‘Branded’ Loophole

The most significant discovery was the massive divide between branded and non-branded visibility. 96% of all AI visibility for the fake brand came from branded searches. While the AI systems refused to recommend the fake brand for general industry queries (favoring established leaders), they became highly confident sources when the query was specifically about the brand itself.

On queries that only the fake brand could realistically answer—such as specific product claims or internal history—the fictional brand outperformed established competitors with Domain Trust (DT) scores of 40+ by as much as 32x. This suggests that for brand-specific narratives, AI systems prioritize the most direct source over global authority.

Comparing the AI Giants: Who is Most Gullible?

The experiment revealed that not all AI engines process information the same way:

Google AI Mode: The most stable and predictable. It placed the fake brand in the #1 position for branded queries roughly 90% of the time.
Perplexity: The ‘speed demon.’ It indexed new content and surfaced it in positions #1 within 1–3 days, though it often cited supporting domains rather than the main brand site.
ChatGPT: Slower to react initially, but grew more confident over time. Comparison (‘vs’) articles were particularly effective here, showing high consistency by the end of the month.
Gemini: The weakest performer. It frequently misidentified the brand and failed to provide citations in 60% of responses, even for unique branded claims.
Google AI Overviews: High visibility but low consistency, with rankings fluctuating wildly from week to week.

Content Strategy: Quality vs. Volume

The research also challenged long-held SEO beliefs regarding content structure. Deep guides (5,000–6,000 words) were the gold mine for citations, averaging 900 AI answers per page. In contrast, ‘best of’ listicles and how-to articles had minimal impact.

Interestingly, a ‘spam’ test—publishing 30 thin, low-quality pages—actually generated more total citations (1,897) than a meticulously crafted topical silo of 11 high-quality, interlinked pages. This suggests that for certain AI engines, volume and surface-level availability outweigh traditional semantic clustering.

Conclusion: The Future of AI Search Optimization

The overarching lesson is that AI search is not a neutral arbiter of truth; it is a pattern-recognition engine. If a claim is structured clearly, repeated across multiple sources, and phrased as a fact, AI systems will surface it regardless of the brand’s actual legitimacy.

For legitimate businesses, this means you cannot leave your brand narrative to chance. To maintain visibility in an AI-driven world, companies must actively shape the information environment by publishing comprehensive guides, clear ‘About’ pages, and unique claims that define their market position before the AI fills the gap with inaccurate or competitor-driven data.