Google has been slowly rolling out its experimental AI-powered Search Generative Experience, or SGE. Like basically every other development in online marketing and search engine optimization (SEO) in the past few months, this has to do with artificial intelligence and its influence on what you see on the internet.
Google has touted SGE as its future, others are raving about it, and online marketers are predicting that it is going to completely transform how we use and search the internet.
At Myers Freelance, we’re going to short the market here because we think it’s just another AI fad.
What is Google’s New AI Search Tool?
SGE is a new way for Google to respond to search queries. Before, the search engine giant would receive a query and use its hundreds of algorithm factors to determine what would be the most relevant and important sites to link to in the search engine results page (SERP). Those algorithm factors were trade secrets and were constantly being tweaked by Google to maximize performance, though SEO professionals were able to determine that some of the factors were things like:
Now, Google is claiming that its AI features, which had been running in the background for years to enhance the search engine algorithms, would be brought to the forefront with SGE.
Based on demos and tutorials, we can see that the old SERP is giving up lots of ground to content that has been generated by Google’s new SGE, which is pulling information and summarizing it for more easy consumption in an area being called the “AI snapshot.”
Where is it pulling that information from? How well is it summarizing it? There are links to other websites in the “AI snapshot” that allegedly “corroborate” the information that has been generated, but how well do they actually do that?
Before, Google’s algorithms were the “black box” that produced results using methods we couldn’t see. After years of studying what came out of it, SEO pros have a fairly good understanding of how it works, at least in the big picture.
The AI aspect of the new SGE search tool is a brand new black box. We’re back to square one.
Google Claims It Will Be the Future of Search
Questions be damned, Google is insistent that this is the future of search. At the July, 2023, earnings call, Google’s CEO Sundar Pichai said that, “over time, this will just be how search works.”
But for all of the pomp that SGE has been getting, it’s been a very slow rollout. SGE has been in what is essentially beta mode since May, with much of that time being “invite only.”
A Big Question: Is This Really What Google Wants to Get Into?
When you think about it, Google’s business model and its place on the web is actually perfect for the business.
On the one side, they have internet users who input queries and, if they’re dissatisfied with the results, generally don’t get upset at the messenger, Google. Instead, they get disappointed with the websites they see or even upset at themselves for phrasing the query poorly.
On the other side, Google has business owners doing everything that they can to please the search engine so it promotes their websites in the SERP.
Even though Google is core to the process, the blowback that it gets when it fails to meet expectations is relatively muted.
The SGE search function grabs hold of way more of the process. Rather than restraining itself to promoting or demoting good and bad results for the query, Google wants to delve into producing the content it shows?
Two serious issues come to mind immediately that will impair Google’s reputation on both sides of the search equation.
First, poor content is going to alienate the internet users, who can directly attribute the failure to Google. If someone asks SGE a question and is dissatisfied with the answer, Google is responsible, not the site that it promoted in the SERP. Unless and until Google has gotten its SGE feature to produce reliably relevant and important content for queries, taking this responsibility seems to be a big risk.
Second, taking on this AI-generative content responsibility will alienate the website owners, who will not only see SGE undermine their traffic in much the same ways that zero-click searches and featured snippets undermined them, but will also have concerns about SGE’s AI programs using the content for their websites to answer users’ questions without visiting their site.
Both sides will be – or at least, should be – perturbed by the new black box that is this AI-generative process. Can users trust what it produces? Does it wisely choose and then effectively summarize “corroborating” websites?
As more data comes out, we’ll have a better idea of how well it performs. But until that happens there’s very little to go on.
AI Has Failed to Meet Expectations Twice, Already
It’s also important to note that AI-powered search engines are not the first uses of artificial intelligence in the online field. We’ve also seen AI-generated images and AI-generated written content, the latter mainly in the form of ChatGPT.
Both uses of AI were highly touted, with extremely ambitious expectations.
AI-generated images have not delivered. It was supposed to break the online marketing game by giving everyone access to quality, affordable, quickly producible imagery that would tank the artistic market.
But it still takes AI artists multiple working days to create quality artwork from start to finish.
AI-generated content took huge leaps forward from earlier iterations, but it’s still failing to deliver quality content. Because it relies heavily on text prediction, AI-generated written content is, obviously, highly predictable. We’ve found that 500 words is about when you can tell that you’re reading AI-generated content – it’s just that boring. It also hallucinates by making up facts and narrating them, quite confidently, as the truth, so you have to fact-check what AI-content producing programs spit out.
While content-generating AI programs have been good enough to throw the legal content writing industry into turmoil, they’re nowhere near as good as even a below average human legal writer. That’s why we’ve predicted that it will not have the staying power in the legal content writing field, at the very least, and will go the way of AI-generated images, though only after creating lots of suspicion between freelance legal writers and their clients.
From these prior iterations of artificial intelligence tools, we can see that making AI programs work well is very difficult.
A Failure Now Would Imperil Google’s Market Share
If SGE’s artificial intelligence program isn’t up to snuff, it would be a serious threat to Google’s market share, especially if it is rolled out as a replacement of its traditional search engine rather than as an alternative to it.
We stopped tracking search engine market share over four years ago because there was so little movement: Google is the big player with around 90 percent of the market.
When you have a stranglehold on the market, there is little reason to innovate. The wiser move would be to respond to innovation. Sit back, own the market, play with AI search internally, but only react if Bing or another “competitor” rolls out an AI-powered search engine that gains any traction. If Bing’s fails, at ease once again.
When you have 90 percent of the market, why risk anything?
Maybe that is the thought behind such a slow walk-out of SGE. It would make sense.
Google is Already Getting Sued for Being Too Big
There’s also the fact that Google is facing a huge antitrust trial for abusing its monopoly on internet searching and advertising. SGE would only increase Google’s role in this process.
We would expect to see Google delay the full implementation of its SGE features until the outcome of the trial is known – which could stretch for years if appeals are made.
If Google loses, we’d imagine that SGE would get shelved – and that Google and other large companies would publicly lament the setback as a grave loss to humanity caused by the renewed antitrust enforcement that they don’t like.