Chris Young
Take a deep dive into the impact of AI and ChatGPT on the future of SEO by examining its strengths, shortcomings, and potential for content creation.
On November 20th, 2022, ChatGPT 3.5 was released to the world as part of an open-source natural language processing experiment. Within less than a week, the platform had grown to over 1 million users, causing the ensuing crash and throttling of the entire system. Since then, the technology has grown and scaled, and reached a level of notoriety that few people or projects can equal. It has quickly escaped traditional marketing and computer science/neural network circles and become part of the international discussion of AI and its place in our world. As is the case with many things these days, if you ask ten different people for their opinions of ChatGPT, you'll likely get eleven different answers.
It’s the latest and greatest! It’s ideologically possessed! It’s crummy and doesn’t provide accurate information! It’s coming for our jobs!
Well, maybe? It might be all of these things or it may be none of these things. What almost nobody is saying right now, however, is "let's test it and wait to see what the possible applications are." Currently, AI and ChatGPT are going through what can only be described as an intense beta-testing phase. Many marketers have flocked to ChatGPT to see just what it can do and what it means for their companies. It certainly is impressive, but will it be able to spit out long-form content that's ready for the printing press sans editing? That may be a stretch. However, the answer to how AI will affect the future of SEO likely lies in the middle ground between "total content creation" and "useless gimmick."
Artificial intelligence (AI) is a complex binomial tree exposed to unimaginably large data sets. By analyzing the "past performance" of existing literature, AI attempts to provide "future results" that are useful and coherent to humans. Within search engines, AI has been used for several years to help curate SERPs. Google, for example, began using AI to modify search results based on bounce rates, time on page, and other KPIs that indicate superior search engine results.
Up until now, AI has mostly been a proprietary tool used by search engines to improve search performance. So, what happens when the opposite—open-source user-facing AI—makes its way to the market? And what does that mean for SEO?
First, low-tier content creation has been democratized. Anyone can input a series of questions to ChatGPT and paste the results together to create a patchwork of content. However, even with the improvements in ChatGPT4, our experience shows that the content created still falls short when compared to human-generated content. The outputs of ChatGPT lack the spirit, or "voice," that copywriters infuse into their writing, and the cadence feels more like a mechanical ride rather than a graceful, living experience. ChatGPT and other AI cannot produce majestic, equestrian similes or delight readers with their writing.
This means two things for SEO. First, since generic content is now accessible to all with almost no marginal cost, it becomes less valuable. Therefore, SEO content needs to increase in quality to rank at the same level. Second, for writers who can only produce generic-quality content, AI and ChatGPT may actually be competition. However, writers with greater talent (and the companies that employ them) have an opportunity to differentiate the user experience with superior content quality.
At best, AI-produced content can inform the reader, but only to the extent that its original dataset is accurate and comports with objective reality. This brings another issue to the surface–ChatGPT doesn’t have a single point of truth, and frequently makes factual errors in the content it creates. While it has access to a broad spectrum of information, it cannot weigh one fact against another and produce a coherent conclusion.
If and when SEOs use ChatGPT to create content, they must be diligent in checking that everything is actually factual and correct. Nothing will discredit your company faster than posting obtusely untrue information that anyone with minor expertise would recognize.
Finally, ChatGPT and other AI cannot work through multifactorial analyses required when creating an SEO strategy. Even to the extent that ChatGPT can write an article, it cannot explain what the content strategy is, nor can it create any justification for why a company should create a given piece of content. For the time being, SEO experts are necessary to guide SEO strategies and make sure they’re properly executed.
Much of the work done by an SEO is characterized by low-yield and high-intensity ideation sessions. This often looks like coming up with ideas for content pieces, outlines, writing meta descriptions, and more. This often takes an outsized amount of time and it represents a minute share of value produced. ChatGPT, because of the data set it works off, has incredible potential to help with the ideation process and can help writers and content creators overcome the dreaded writer's block. It’s important for writers and SEOs to recognize that mature ideas won’t flow forth from ChatGPT; any ideas still need to be vetted and polished by the SEO. This is merely a tool to help with ideation. However, ChatGPT can produce 10 article ideas in less than 30 seconds. It can produce a whole content brief or article outline with subpoints in less than a minute. This is a game changer for content creation. Yes, it still needs to be verified, polished and checked for factual errors. However this entire process can frequently be reduced from an hour or two down to five to ten minutes. If you’re not ChatGPT to ideate topics, it’s very likely you’re leaving hours of billable time on the table.
Bard is the latest and perhaps most hurried development by Google. The project was launched with all earnestness and none of the normal pomp and circumstance that normally accompanies Google’s releases. This platform has likely been in development for a while, however it was expedited by chatGPT’s arrival to the market and rightfully so. So many people rushed to ChatGPT during its first week of notoriety that the servers crashed–obviously there’s a market opportunity.
However, Bard isn’t attempting to copy ChatGPT. Rather, Bard is more oriented towards providing a great user experience and answering quick questions in a coherent way. ChatGPT on the other hand is a bona fide attempt at creating a non-human apparatus for novel content creation. It’s helpful to think of Bard as Google’s concierge desk, and chatGPT as if it were a combination of Wikipedia and your favorite mainstream journalist.
The main risk that comes with using ChatGPT and other AI is relying on it for things that it just cannot do. Open AI (the organization that created ChatGPT) is totally transparent with the limitations that come with using the ChatGPT software.
The first area in which ChatGPT is limited is in its knowledge. Some of the answers that it gives to different prompts are just wrong and, in some cases, there is no data set for it to pull from. For example, ChatGPT also doesn’t possess any particular knowledge about the state of a given company–there’s no data set for it to use. So it may be able to discuss, for example, the automotive industry. However ChatGPT couldn’t produce a press release or a thought leadership article for a car manufacturer.
AI, on a broader level, also has the potential to be used in malicious ways like creating fake news or promoting libelous or slanderous information. Additionally, programs like ChatGPT cannot even cite their sources.
This walled garden of primary data makes verifying specific sources a nearly impossible task. When applied to hot button issues, this can easily create a no-win scenario for risk averse companies. As a result, they may decide to simply remain silent, or worse, pick the side with prevailing cultural winds despite evidence to the contrary.
This leads to two different problems, one moral and one transactional. The first is that if (and when) AI gets to the level where humans are no longer directing the outputs, there is nobody to hold accountable for otherwise illegal or immoral activity–the AI did it! Furthermore the lack of transparency in the source of information or “showing the work” of how an AI reaches one conclusion over the other can leave users in an epistemological conundrum. How are we as users supposed to know what is true and what isn’t? Even worse, after several years of AI-produced content, a vicious cycle of AI referencing AI may emerge. The time and energy needed to unravel this web could be gargantuan.
In terms of SEO, all of this translates into content creation that is of questionable quality and veracity when left unchecked by human eyes. Especially when Google and other search engines value “human-first” content (which by necessity must be true and high quality), this presents a problem for AI-generated content. SEOs are now tasked with ensuring that any content produced by an AI is both human-oriented and as true, precise, and accurate as it can be. Otherwise the content may be subject to ranking penalization by search engines.
Even with the general utility of natural language processing from AI, it’s likely that users will continue using search engine results as their main research avenue. This is driven by one main reason: many people who go to search engines are engaging in research and seek multiple perspectives rather than asking to be told what to do. ChatGPT and other AI platforms will tell users what to do/what is true whereas the first SERP turns out 20 different answers to a user’s query which can then be evaluated based on the user’s criteria for truth.
Each page that ranks has different nuances. This panoply of view points ultimately provides the users with more options, a better problem-solving experience, and a greater data set for decision making. Current AI models are monolithic in comparison. In fact, we might be able to say that AI use and problem complexity have an inverse relationship; as a problem’s complexity increases, reliance on pure AI input decreases. This might be demonstrated by someone searching for how to remove wine stains from a t-shirt (total reliance on the AI’s answer) and a postgraduate student doing research for their thesis (perhaps zero reliance on AI).
As a result, high-quality SEO oriented content will still have a place in any company that wants to pursue organic content creation. In fact, the emergence of generic AI-produced content will likely create a new way for companies to differentiate themselves with high quality, thought-provoking content.
The future of SEO with AI will likely be characterized by a symbiotic relationship that allows SEOs and content creators to develop SEO content with a greater degree of alacrity. Foundational work for SEO content creation can be expedited by leveraging AI to help with the ideation process. Even some types of content like product reviews can be somewhat automated by AI if this information was included in the original data set. This can save hours of scrolling through Amazon reviews, reddit boards and more trying to find the right information. In essence, AI should be one of many tools that exist in an SEOs toolbox, right along SEMRush, AHrefs, and Screaming Frog.
While AI platforms appear to create a shortcut in ideation and other foundational work, AI like ChatGPT still aren’t talented enough to create anything other than generic content. As a result, using unedited AI-content should be avoided at all costs; humans will be turned off by it and Google will notice. Any content with “soul”--that intangible quality that makes something literature rather than just information–must be written by human hands.
There may be a future where AI can dominate human language and emulate human spirit to a degree where human-generated content and AI-generated content are indistinguishable. While this would solve some problems, it would likely open a Pandora's box of other, greater problems, and not just in SEO. Most likely, AI will continue to advance in a capacity that enhances human capabilities, rather than in a way that replaces humans.
For those marketers that choose to include AI among their suite of tools (and they should include it), they must also develop a new skillset. Prior to the industrial revolution, humanity had to rely on skills that were based on individual strength–hammering, plowing, pushing and pulling. Then heavy machinery came along and these old skills based on strength gave way to new skills like the ability to maneuver joysticks, turn steering wheels and control the velocity of internal combustion engines. The changes that affected humanity during the first industrial revolution are now the same ones that will affect intellectual workers; we must adapt to a different set of skills based around working with AI to write rather than just writing ourselves.
With the current user interface that ChatGPT uses, the most important NEW skill an SEO can possess is the ability to ask good questions and write good prompts. The quality of an AI output is directly related to the quality of the prompt created by the human. In addition to getting better at working directly with artificial intelligence, marketers must also learn how to incorporate into their current suite of SOPs and processes. This will look different for each company. However best practices will soon emerge with regards to the proper use and context of AI platforms.
So what exactly is AI and what does it mean for SEO? Well, It’s neither a bird, nor a plane, nor is it superman. It’s merely a tool, just like many other tools currently on the market, that can increase the value of SEO to different companies. We’re in the very early days of understanding the significance of AI and the exact ways in which it can be used. However, as a tool, AI must be part of a small part of the SEO strategy rather than the entire strategy. Sorry, you can’t fire your CMO quite yet!
Tectonic shifts in technology like AI and ChatGPT can leave businesses and marketing teams with fundamental questions about how to move forward in the new environment. Be More Digital has been on the forefront of marketing advancements and natural language processing AI (like chat GPT) and can help you understand how to use it in your marketing plan. Additionally, in a world where generic content is exploding, real content with an engaging voice will become crucial to attracting visitors. Be More Digital can help maximize engaging content and out-compete your competitors by creating SEO-optimized human-written content. For more information, contact us or visit our websites for a full list of service offerings.
.png)
