ChatGPT and SEO

In 2024, ChatGPT became one of the most widely discussed tools in the digital marketing space. Every SEO professional, agency, and business owner started asking the same question: can ChatGPT-generated content rank on Google?

At Gilead Digital, we decided not to guess. We ran four distinct tests over several months to understand exactly how AI-generated content performs in real search conditions. This article documents our findings, including what worked, what failed, and the hybrid approach that produced the best results.

When ChatGPT launched, the marketing industry was split. Some believed AI would replace content writers entirely. Others dismissed it as a tool that could never match human creativity and domain expertise.

We took a practical approach. Instead of relying on opinions, we set up controlled experiments with real client websites and our own properties. The goal was simple: determine whether ChatGPT-generated content can rank on Google, and if so, under what conditions.

For our first experiment, we created 26 location-specific pages for a client using content generated entirely by ChatGPT. Each page was supplemented with relevant images to ensure a complete user experience.

The initial results were promising. Within the first week, several pages started appearing in search results and showed decent ranking positions. However, the success was short-lived.

After 2 to 3 weeks, every single page lost its rankings. Google appeared to have re-evaluated the content and determined it did not provide sufficient unique value. The pages dropped completely from the search results.

Key takeaway: Purely AI-generated content may achieve temporary visibility, but Google's algorithms can identify and devalue it over time.

To confirm our findings from Test 1, we replicated the same approach with a different client. We used ChatGPT to generate location pages following a similar structure and publishing process.

The outcome was identical. The pages initially gained visibility in search results, but within 2 to 3 weeks, the rankings disappeared. This confirmed that the pattern we observed in Test 1 was not an isolated incident.

Key takeaway: The ranking loss pattern is consistent and repeatable, suggesting that Google has mechanisms to detect and penalise purely AI-generated content at scale.

For the third test, we applied ChatGPT-generated content to our own website. We wanted to see if the same pattern held when the domain had existing authority and trust signals.

The result was disappointing. The AI-generated content did not rank well, even on our own domain. It was only after we manually rewrote the content — adding our own expertise, unique insights, and a more natural writing style — that the pages began to achieve rankings.

Key takeaway: Domain authority alone is not enough to make AI-generated content rank. The content itself needs to demonstrate genuine expertise and originality.

Based on our learnings from the first three tests, we tried a hybrid approach. We used ChatGPT to generate approximately half of the content — primarily for structure, outlines, and initial drafts. The remaining half was written from our own experience and expertise.

This combination produced the best results. Several pages achieved top 3 rankings for keywords with difficulty below 25. The hybrid content maintained its rankings over time, unlike the purely AI-generated pages.

What made the difference:

• ChatGPT provided a solid structural foundation and saved research time
• Human editing added original insights, real examples, and a natural voice
• The content felt authentic because it combined AI efficiency with genuine expertise
• We targeted keywords with manageable competition levels

Key takeaway: The most effective approach is to use ChatGPT as a starting point, not as the final product. Human oversight and original thinking are essential for sustainable rankings.

After months of testing, here are the core insights we gathered:

• 100% AI content is risky: Google can detect and devalue purely AI-generated content, often within 2 to 3 weeks of initial indexing
• Temporary rankings are misleading: Initial visibility does not guarantee long-term performance. Always monitor rankings over a 4 to 6 week period before drawing conclusions
• Domain authority does not compensate: Even established domains cannot rely on AI content alone to rank. Quality and originality matter at every level
• Hybrid content performs best: Combining AI-generated structure with human expertise, original data, and authentic voice produces sustainable results
• Keyword difficulty matters: The hybrid approach worked best for keywords with lower competition. Highly competitive terms still require deeply original, expert-level content

ChatGPT is a powerful tool for SEO when used correctly. It excels at creating outlines, generating first drafts, and helping with research. But it should not replace the human element in content creation.

The winning formula we discovered is straightforward: let AI handle the heavy lifting of initial drafts and structure, then layer in your expertise, unique perspectives, and real-world examples. This approach saves time while maintaining the quality standards that Google rewards.

If you are experimenting with AI content for SEO, we encourage you to run your own tests. The results may vary depending on your industry, competition level, and content quality. But based on our experience, the hybrid approach consistently delivers the best outcomes.

For expert guidance on integrating AI into your SEO strategy, reach out to Gilead Digital. We help businesses build content strategies that leverage the best of AI and human expertise.