How Earned Media Influences AI Citations and Brand Visibility

Groundbreaking Study Reveals How AI Chooses Its Citations

Miami, July 23, 2025 — A new study from Muck Rack, a leading PR software provider, has uncovered how generative AI models decide which sources to cite. By analyzing millions of AI-generated citations from hundreds of thousands of prompts, the research reveals the direct impact media coverage has on AI outputs—and why this matters for PR and communications professionals.

At its core, the study answers a critical question: Does media coverage shape what AI says? The findings confirm that it does.

Key Takeaways from the Study

• Citations actively change AI responses. Testing with large language models (LLMs) like ChatGPT and Google Gemini showed that when citations are included, AI outputs shift significantly. This means cited sources aren’t just decorative—they actively ground the AI’s answers with current, relevant information.
• Earned media is crucial. Over 95% of AI citations come from unpaid media, and 85% of those are earned sources. About 25% come from journalistic outlets. Half of all AI responses contained at least one earned media citation, highlighting the importance of PR strategies focused on earned media for enhancing AI visibility.

What Drives AI to Cite a Brand?

Three main factors influence whether a brand gets cited in AI-generated content:

• Recency: Fresh content is favored, especially by OpenAI models, for queries related to current events, opinions, or trending topics.
• Query framing: Questions seeking advice or opinions prompt more dynamic citations. Encyclopedic or fact-based queries tend to rely on older, static training data instead.
• Outlet authority: High-authority sources like Reuters, Axios, Financial Times, AP, TIME, Forbes, NPR, and CNN are often cited. However, niche sites such as Good Housekeeping and Investopedia also appear frequently.

What This Means for Communications Teams

PR strategies must adapt to this new AI-driven layer of brand visibility. Success metrics should now include whether your brand is cited in AI responses—not just traditional media placements or backlinks.

To improve your chances of being cited, prioritize coverage in high-authority outlets and time your media releases to align with AI’s preference for recent content.

As AI becomes a primary way people access information, your PR approach needs to address how algorithms interpret and present your brand—not just how humans see it.

The study underlines a clear point: the stories you place, and where and when you place them, influence how AI understands your brand.

Study Methodology

The research involved running hundreds of thousands of prompts across multiple web-enabled language models during July 2025. These included ChatGPT (versions ‘4o’ and ‘4o-mini’), Gemini (‘flash’ and ‘pro’), and Claude (‘sonnet’ and ‘haiku’).

Prompts covered a wide range of industries and topics, sometimes mentioning companies directly and sometimes not. The goal was to quantify how often AI cites sources, what types, and the role of earned media in its outputs.

About Generative Pulse

Generative Pulse helps PR and communications teams track how their brands appear in AI-generated search results. Its focus on Generative Engine Optimization (GEO) reveals which journalists and outlets influence language models like ChatGPT. Integrated into Muck Rack’s PR platform and backed by significant funding, Generative Pulse equips communication professionals with the insights to manage brand visibility in AI-driven search environments. Learn more at generativepulse.ai.

About Muck Rack

Muck Rack offers award-winning PR software that combines media database, monitoring, and reporting tools for seamless collaboration, pitching, and measurement. Serving over 5,000 companies worldwide, it provides the most accurate data directly from journalists. Thousands of journalists also use Muck Rack’s free tools to showcase their work and measure story impact. More details at muckrack.com.