The rise of large language models (LLMs) is transforming the consumption and production of digital content, from blogs to search engines and traditional internet sources. We’re witnessing a future where search engines aren’t the only practical way to navigate the wide, dense forests of online information — and where humans are no longer the sole creators of content.
From Google testing an AI news article writing bot to AI-powered search engine optimization (SEO) tools that can optimize reams of content in milliseconds, as well as bloggers automating their blogs with ChatGPT, AI is profoundly changing the digital content landscape.
So what does this AI revolution mean? Will this boost in AI platforms reduce the prominence of search engines and other content platforms? How will these changes impact traditional marketing efforts? Read on to find out.
Generative AI: A New Way to Find Answers
The internet has always been a place to find answers to questions — and, until recently, we’ve done that through the search engine.
Since its inception, Google, like other search engines, has strived to provide valuable responses to natural-language questions. The search engine Ask Jeeves, launched in 1996, was entirely based on the premise that users ask it questions.
But asking questions is only half of the equation: The more crucial role of a search engine is to generate relevant answers or results.
For example, if you ask Google a question, you’ll receive a wide range of results. If your question has a clear answer, you’ll see direct answers first, followed by results below that get progressively less pertinent.
But when you ask questions to an LLM such as ChatGPT, the format of the results is completely different. At its best, the output from an LLM is remarkably human-like, reflecting a nuanced understanding of the questions asked. Its answers stem from its training data, an immense quantity of text data on the internet. It has access to much of the same data as a search engine but delivers its response to a question as full sentences and paragraphs, rather than an extensive list of links.
We regularly use search engines to find answers to questions, but they’re actually designed to help us find relevant websites — we often have to do some further digging to find answers. LLMs like ChatGPT, on the other hand, do the legwork for you. So, it’s easy to see how tools leveraging LLMs might replace search engines in many situations — at least until search engines incorporate these capabilities.
That said, there’s at least one reason ChatGPT hasn’t taken Google’s top spot yet. As OpenAI states at the bottom of the ChatGPT homepage, “ChatGPT may produce inaccurate information about people, places, or facts.” While LLMs can offer a range of responses, they sometimes “hallucinate” facts. They’ve been known to name non-existent towns and cities, make up quotes from historical figures, and invent statistics to support their claims. But the biggest problem with LLMs stems from the fact that despite their inaccuracy, they speak with an eloquence that inspires confidence.
Accuracy in search results has been an issue since search engines existed. However, with a traditional web search, the onus is on the user to verify the reliability of the results. Google aims to prioritize reliable websites and has significantly improved this over the years, yet it still provides links to websites that contain inaccurate information. ChatGPT, on the other hand, gives authoritative, accurate-sounding answers that are often correct but can also contain inaccuracies.
LLMs still have a long way to go before replacing search engines — if that’s ever likely. However, they’re already having a significant impact when it comes to generating written content and searching for information.
The Impact of AI-Generated Content on Search Engines
Historically, Google has penalized automated content for being poor quality and spammy. In the early days of the internet, black-hat SEO tactics often involved automatically generating, or “spinning,” content to populate websites, aiming to manipulate rankings without investing in genuine content creation. Algorithm changes from Google, especially the Panda update, targeted such practices, pushing websites with poor content down in search results.
The landscape has shifted dramatically with the rise of advanced AI content creation. Detecting AI-generated content is extremely difficult, if not impossible, so it’s not surprising that Google’s policies on automated content have evolved. After all, the search giant is primarily concerned with ensuring users can find valuable, accurate information — it only penalized AI-generated content because it was poor quality.
For anyone involved in producing search-engine-optimized content, this has profound implications. Marketers can now use LLMs like GPT-4 as another tool in their SEO arsenal, as long as the resulting content adheres to Google’s guidelines. Unless the content is plagiarized, its origins don’t matter. What’s important is that the content aligns with Google’s E-A-T framework and demonstrates “expertise, authoritativeness, and trustworthiness.”
Navigating a New Era of Content Creation and Consumption
As we’ve already discussed, AI is reshaping how we consume content. Tools like ChatGPT are already invaluable for creating SEO content quickly. However, AI is unlikely to entirely overshadow traditional search engines or publishing for several reasons, including those below.
Training Data
LLMs like GPT rely on their training data — which consists of existing, human-written texts — to generate answers. At the time of writing, ChatGPT’s knowledge only extends up to its last training date in September 2021. It will inevitably need to train on new, original content or become progressively less useful as its training data becomes outdated.
Traditional publishing, therefore, will remain important even as the use of AI tools becomes more widespread. Brands, products, and novel ideas still need to be introduced to these models — just like they do to humans.
Fact-Checking
Even if LLMs evolve to attain superhuman accuracy, there will still be the necessity and desire to locate primary sources and perform searches outside LLMs. Users often seek to trace information back to its origin to gain a more comprehensive understanding or to verify the authenticity of the content generated by LLMs.
In this regard, search engines and “traditional” content remain indispensable as they provide a pathway to original sources, enabling a deeper, more authentic exploration of topics.
The Human Perspective
As the internet becomes populated with more AI-generated content, an ever-greater premium is placed on genuine human perspectives. We can already see this in Google’s recent addition to the E-A-T framework: It’s now E-E-A-T, with the extra E standing for “experience.” LLMs will never be able to claim firsthand experience like a human can, so this is one way that Google still privileges human-created content.
Originality
Originality is related to the previous point. While AI bots like ChatGPT may sound fluent and authoritative, their content is based entirely on pre-existing written content. They can’t generate new ideas — their capabilities are limited to rehashing existing ones. Relying solely on highly derivative, AI-generated content is unlikely to be effective marketing in the long run. If you need something new, you need a human.
Next Steps
AI’s impact on content creation and consumption has already been transformative. Yet, amid this revolution, traditional publishing, search engines, and the enduring power of human perspectives persist.
As we embrace the possibilities of AI-generated content, we must respect the authenticity and richness of human-made content. The most optimistic vision for AI’s future is that machine intelligence’s tremendous power becomes an aid, rather than a replacement, for human creativity.
Looking for a partner to help you navigate the ever-evolving world of content marketing? Reach out to ContentLab today.