Is it true that LLMs prioritise established media and press releases?
We’ve entered the AI era, where large language models (LLMs) like ChatGPT and Perplexity are key tools for finding information. The big question everyone is asking now is how this affects their communication. In this article, we examine whether it’s true that LLMs favor established media channels and press releases when generating answers. If so, what does it mean for us working in PR and communication?
LLMs Are Trained on Content from Major News Publishers
Studies show that LLMs prefer content from major news publishers because that’s how they are trained. For example, OpenAI’s GPT series was trained using the Common Crawl, which includes material from major news sites such as the BBC and The New York Times. This media-heavy training data makes LLMs mirror the style and perspective of these sources.
The general trust in these sites is also important for LLMs – something built up by a large number of readers and human interaction. This means that LLMs are more likely to reproduce widely accepted information from the press and are less likely to introduce viewpoints that deviate from media consensus.
For someone working in PR and communication, this means that getting your story into established media is more important than ever to make your voice heard and influential.
Not Without Challenges: Bigger Isn’t Always Better
As mentioned, it has been proven that LLMs prefer using established media sources when generating content. An example is the 2024 Taiwan elections, when Bing Chat primarily used English-language news articles to generate answers asked by LLMs, overshadowing local voices.
Another study shows that LLMs often rely on press releases or easily accessible sources rather than seeking varied or investigative perspectives. To increase the chance of being noticed, it’s essential to create clear and accessible press releases that LLMs can read, preferably ones that established media outlets also pick up.
Risk of Local Journalism Disappearing in AI’s Shadow
At the same time, there should be efforts to highlight more varied and investigative sources to broaden perspectives. There is concern that LLMs may lead to only a few media channels dominating, risking local and independent journalism being overshadowed. Additionally, concerns about the lack of source transparency in AI-driven search results and content creation persist, as users may not always be aware of where the information originates or what biases it may contain.
Conclusion
In the AI era, where LLMs like ChatGPT and Perplexity are increasingly used for information retrieval, content from traditional media and press releases is currently favored. This is because LLMs are trained on data from established news sources, which means they tend to draw information from these outlets or from sources that are structured in a similar way.
This is what the landscape looks like right now, but everything is changing quickly these days. As technology and media consumption habits continue to evolve, it’s more important than ever for PR and communication professionals to write clear, accessible press releases that established media can easily pick up.
At Mynewsdesk, we’re here to help you navigate this shifting environment. We can support you with everything from crafting compelling press releases to connecting with the right journalists who are interested in your story. We also offer monitoring services, making it easy for you to track what’s being said about your brand, spot trends in your industry, and identify which media outlets are most likely to cover stories like yours.