How Will Generative AI Impact SEO?

When ChatGPT debuted in November 2022, it was heralded as the biggest disruption to SEO since the industry was created more than twenty years ago. Some said it was the end of SEO as we know it. The truth is that ChatGPT and other generative AI products have had very little effect on SEO so far. 

As this chart from Similarweb shows, Google’s share of search traffic worldwide has remained consistent at 91.05% over the last 12 months, down to ⅒ of a percentage point, despite the massive wave of interest in ChatGPT and lack of meaningful response from Google. 

Still, generative AI seems poised to change SEO in a few ways over the next few years. For example, on the searcher side, there will be more “no-click” engagements where visitors remain in Google’s ecosystem conversing with a chatbot. And on the marketing side, content creators will use generative AI tools to research better, summarizing complex web pages and generating graphics quickly, raising the bar for content.

In this piece, we break down all the ways generative AI will be changing SEO – as well as the ways it won’t be – from a sober, big-picture perspective.

Impact of Generative AI on SEO, Summarized 

After a great deal of discussion on how SEO will change with increased usage of generative AI tools, our team agreed on the following: 

Why Generative AI Won’t Replace SEO

Before we go into detail about why SEO won’t be overtaken by generative AI, let’s begin with a question: Do you use ChatGPT to make important decisions about how to spend your time or money? Or do you mainly play with it, ask it interesting questions, and watch it compose astonishingly humanlike answers?

If you don’t yet trust chatbots to make important decisions for you, you’re already getting at the first reason why SEO isn’t being replaced anytime soon. 

People Don’t Trust Generative AI 

As of today, we don’t trust generative AI enough to use it as a replacement for Google. Why? 

1. It is often wrong. In the screenshot below, ChatGPT claims that the scientific name for the hedgehog, “Erinaceus Europaeus” is the same as its English name “hedgehog,” which is nonsensical.

2. It does not have up-to-date information on many subjects. In the screenshot below, ChatGPT claims that the largest volcano eruption ever recorded was in 1815; it was not aware that a larger eruption occurred in January 2022.

3. It does not give specific, helpful opinions to commercially-oriented questions. Below, Google Bard responds to “What is the best SEO agency?” with vague criteria that bring the searcher no closer to the answer they were seeking.

Given these limitations, generative AI is ill-prepared to replace Google’s core function of helping people make decisions. And if it can’t do that, its chances of displacing SEO as a marketing channel are low. After all, the fact that we trust Google to counsel us when the stakes are high is the core reason SEO is such a powerful channel in the first place.

Of course, not every Google query is high stakes; a substantial percentage of them are in the realm of basic research, e.g. “what is a parquet floor?” – the very types of queries that generative AI is best suited to answer. Yet losing users to chatbots for those types of searches isn’t very impactful to Google. Basic research queries have always been a loss leader for them, as few companies want to advertise around searches with low buying intent. The lion’s share of investment in both SEO and Paid Search is around transactional keywords like “interior designer brooklyn” or “best parquet flooring brands”. Of course, Google would rather you stay in their ecosystem for all information discovery, but none of us are ever too far from it between Google Search, Gmail, Google Maps, Chrome, Android, and Drive.

Generative AI Can’t Write Expert Content

Marketers hoping to use generative AI products like ChatGPT or Google Bard to write content for them encounter an interesting phenomenon: the content is acceptable, but not very convincing. It has an even-handed tone that conveys neutrality at the expense of persuasiveness. 

As an example, compare the two answers below to the question “What are the main differences between selling a business to private equity vs a strategic acquirer?” The first is from Google Bard and the second from an M&A expert.

Google Bard Answer

Expert Answer

Marketers that rely too heavily on generative AI to produce SEO content will find that their content is lackluster, and more importantly, of equal quality to the many other marketers in their industry doing the same thing. In the meritocracy of Google, where there is only one #1 result, standing out is essential. Thus, genuine thought leadership written by subject matter experts will continue to beat out AI-produced content in the realm of SEO and the larger context of earning the attention of potential buyers.

Legal & Competitive Barriers Will Limit The Effectiveness of Generative AI

A common retort to the assertion that generative AI is inaccurate or less trustworthy than a human expert is “It’s only going to get better.” While true, this idea ignores the inevitable legal and competitive resistance the technology will face. Simply put, IP holders and content creators will not give away their content for free.

For years, Google has maintained an environment where website owners voluntarily share their knowledge publicly in exchange for exposure in the search results. If generative AI companies use this knowledge to feed their chatbots without giving the same exposure, website owners will lose their incentive and stop creating free content. ChatGPT avoids this problem by utilizing training data that’s out of date anyway. Bing AI, on the other hand, harvests newly-created content from websites, and in doing so breaks the tacit agreement website owners made with Google in the 2010s that incentivized this content creation in the first place. 

If search engines continue down the path of feeding chatbots user-created content without consent from those users, it will disincentivize them from creating it. Larger IP holders such as Dow Jones, Reddit, and Twitter have already demanded that Microsoft cease scraping their data to train its chatbots.

If large language models could access all the knowledge in existence and people kept training them, we could achieve a kind of ‘generative AI utopia’ where information on every subject is available when we need it and can be combined in novel ways to expand our understanding of the world. But unfortunately, no such proposal has been made by OpenAI, Microsoft, Google, or any other player in the space. Instead, commercial interests have driven the rise of generative AI from the start, and free market competition will be the result. 

In the near future, we will see more companies gating their content, preventing LLMs from accessing their most up-to-date ideas so they themselves can benefit from it. One can imagine a future where SaaS businesses compete to create the best-trained chatbot in fields such as finance, health, and marketing. We’ll see personalized chatbots for various needs such as a meditation chatbot from Deepak Chopra or a motivational chatbot from Tony Robbins. But each of these services will be its own business.

These legal and competitive obstacles will fragment AI services, collectively taking attention from Google Search, but again, failing to replace it. Searchers will continue to seek out the universal provider of relatively trustworthy information that is Google, and marketers will still need to create content to rank highly in the search results. 

Future-Proofing Your SEO in the Age of Generative AI

The predictions of SEO’s demise at the hands of generative AI remind me of similar predictions I heard when travel websites like Expedia and Travelocity came out in the 2000s. At the time, the accepted wisdom was that travel agencies would be wiped out soon. That wisdom was both right and wrong: The low-end agencies that just booked your flights and hotels did become obsolete. But the high-end ones that offered genuine expertise about where, when, and how to travel not only survived, but are thriving to this day.

SEO will follow a similar pattern; marketers and agencies that create lower-effort content will become irrelevant, as generative AI can do the same. But marketers that create true thought leadership, replete with original ideas and analysis, will be valued more than ever. Particularly in B2B SEO, where decision-makers are searching for answers that have major financial impact on their business, expert guides will always be needed.