Since the introduction of generative AI, large language models (LLMs) have become widespread in search engines. But can we influence their performance through large language model optimization (LLMO) or generative AI optimization (GAIO)? In this article, we will explore the evolving landscape of SEO and the uncertain future of LLM optimization in AI-powered search engines, with insights from data science experts.
What is LLM optimization or generative AI optimization (GAIO)? GAIO aims to help companies position their brands and products in the outputs of leading LLMs, such as GPT and Google Bard, as these models can significantly impact purchase decisions. For example, when searching for the best running shoes for a 96-kilogram runner who runs 20 kilometers per week, Bing Chat suggests brands like Brooks, Saucony, Hoka, and New Balance. Similarly, when searching for safe, family-friendly cars suitable for shopping and travel, brands like Kia, Toyota, Hyundai, and Chevrolet are suggested.
These recommendations from generative AI tools, including Bing Chat, are contextual and rely on neutral secondary sources such as trade magazines, news sites, association and public institution websites, and blogs for recommendations. The frequency of word co-occurrences in the training data determines the statistical probability of their appearance in the output. LLMs, like GPT and Bard, operate based on statistical analysis and utilize semantic spaces represented by vectors to process texts and data.
While LLMs rely more on statistics, they are improving in semantic understanding due to the vast amount of data available. Entities mentioned frequently together in the training data show a high statistical probability of a semantic relationship. The approach of transformer-based natural language processing (NLP) involves transforming natural language into a machine-understandable form to facilitate communication between humans and machines.
So, can the outputs of generative AI be influenced proactively? Data science experts have provided different perspectives on this question. They note that commercial large language models, like GPT and Bard, keep their training data private and use alignment strategies to ensure the AI’s responses remain neutral. Therefore, intentionally influencing the AI’s opinion would require over 50% of the training data to reflect the desired sentiment or flood the internet with posts and texts to be incorporated into the training data.
The complexity and challenges of influencing LLMs through content, PR, and mentions make it an unfeasible approach. The dynamics of network proliferation, time factors, model regularization, feedback loops, and economic costs are obstacles to making a meaningful impact. The identification of reliable sources for training data is also a challenge.
In conclusion, while it may be challenging to influence the outputs of generative AI, there is still the opportunity to optimize for search engines by focusing on PR and marketing efforts. By understanding the dynamics between LLMs, systems like ChatGPT and BARD, and SEO, businesses can better position their products and brands.
At Bridgewell Marketing, we specialize in providing effective SEO services to help businesses improve their online visibility and generate more organic traffic. Contact us here to learn more about how we can optimize your website for search engines and drive better results.
