03 / August 2016, 1st In SEO
The concept of artificial intelligence, or AI, has existed for centuries, even if the phrase itself wasn’t coined until 1956. The idea that humans could create something capable of thought processes similar to, or even superior to, their own, existed with the ancient Greeks and has extended through the millennia. The concept ramped up significantly in the 1950s, though computer memory and construction limitations prevented significant breakthroughs from occurring. Science fiction novels and films began foreseeing an ominous future, and recently, AI applications have started infiltrating our world. More recent years have seen some interest spikes: Deep Blue, a chess-playing supercomputer, defeated Garry Kasparov in 1997, and IBM’s Watson destroyed its human competition in Jeopardy! in 2011. And in 2015, Google announced the arrival of RankBrain as a critical component of its search processing strategy.
But before we throw our hands up in despair and conclude all is lost to our computer overlords, we should step back and look at how AI really works. It can provide tremendous advantages for those employing it, to be sure. From an SEO perspective, though, it may mean less—and more—than many seem to fear. It’s true that we learn a little more every year just how much we don’t know. When we examine what reality currently provides, though, we can begin to assess what AI means to SEO work, both now and in the foreseeable future.
What AI Cannot Do
For movie buffs, it may be important to dispel a few knee-jerk reactions to the notion of artificial intelligence. Skynet is not about to become self-aware. The machines are not going to use humans as a plug-in source of energy, even if we do block out the sun. And Hal is not going to start eliminating crew members. The humans are, in fact, still in charge, and the technology at hand is here to assist us, not to take over.
When we get past sci-fi layers of dread and discomfort, we can look at what artificial intelligence can in fact do. One useful way to analyze it is to consider the different levels of AI that have been theorized. Artificial Narrow Intelligence, or ANI, comprises all of the current AI technology in existence. This allows processing of data at very high speeds to perform specific tasks. Deep Blue, Watson, RankBrain, and high-tech appliances and assembly-line machines can all do this for different kinds of work. For the tasks in which computers and machines are programmed, they can usually operate far more efficiently than a human mind. For this, they have become incredibly useful in different areas of society.
Artificial General Intelligence (AGI) would consist of technology that can function as well as a human mind on every task that humans do. It would move beyond calculations and discrete tasks to perform abstract reasoning, planning, problem solving, and learning quickly from experience. It does not currently exist, and would in fact require a quantum leap in technological development from any ANI currently in place.
Finally, Artificial Super Intelligence (ASI) would perform better than human minds at tasks across the board. It would move beyond logical progressions to be wiser, more creative, and more socially adept than humans, while performing other tasks better as well. This is the level at which science fiction launches into dystopian visions of human slaves subjugated to the whims of mechanized masters. And like AGI, it does not currently exist.
So you might reasonably ask what this has to do with SEO. With AI only at the sunrise of its day of impact on internet search functionality, it matters a great deal. Analyzing ANI capabilities can help us begin to understand present-day impact on SEO functionality and planning. Looking ahead to the development of AI technologies, though, it can provide either comfort or despair for your approach to SEO, depending on your point of view and your methods.
AI in SEO Today
Google’s RankBrain, and the RankNet AI tool that Microsoft began using with Bing in 2007, are AI tools currently employed in internet search. The respective companies use these systems for similar purposes: to help get at the meaning of what users actually want to find. They do this through programming that more efficiently processes search terms entered, but also more effectively uses the words and phrases entered to reach concepts. By analyzing search queries, results, and clicks on the results that occur every day, each tool “learns”—that is, processes and applies the data to generate more effective search results.
This comes into play particularly for long-tail keyword phrases, but can apply to any phrase that means more than the sum of its words. The phrase “Barack Obama’s wife,” for example, is not a search for Barack Obama, but rather for Michelle Obama. Similarly, Donald Trump’s wife”represents a search for Melania Trump, not for Donald Trump. One can envision long-tailed
searches that, absent this parsing ability, provide search results that include pictures of our current president in a sleeveless dress, or the Republican nominee in a bikini. AI, by analyzing and determining the meaning of the respective search phrases, helps save us from that fate.
The goal of the AI search tools is tied to the overall goal toward which Google and Bing strive: connecting the right information to the right people. Some have questioned whether RankBrain is a signal within the algorithm (Google has stated it to be the third-most important signal in its Hummingbird algorithm), or is in fact a tool that processes the algorithm against a subset of search queries.
While AI and machine learning are usually thought of best as high-powered data processing tools that analyze, organize, and respond to data, the understanding of it as a ranking signal is probably most accurate. Besides Google ascribing that nomenclature to it, the analytics that go into determining what a search query means will never be the only, or even most important,aspect of what search results follow. Once Google (or Bing) arrives at the best interpretation of a search phrase’s meaning, it still must identify the search results that best meet the needs of someone searching for that meaning.
This brings us back to keywords and SEO copywriting. The idea that AI renders SEO meaningless ignores the reality that search still must anchor itself to something. When you look to establish high rankings for a particular keyword phrase, you may not need to focus on the exact language, but you still need to include within the post or page words that demonstrate your page is in fact about that concept. And on the flip side, when you provide excellent content, you may find yourself ranking for a search phrase that never appears verbatim on your page.
High-Performing SEO Content
The aspects of great SEO Content, in other words, do not change with the current iteration of AI in web search. Content that is well-written, relevant to the subject matter for which you optimize a search, fresh for readers, and reflective of authority through other signals within the search algorithms will perform well. Literal text that does not possess these attributes, on the other hand, will not do as well as it once did.
This all means that Google has launched further ahead in its never-ending battle against gamesmanship. SEO content providers generally fall into two main camps. The first camp builds its offering to website owners around its ability to trick the system. What initially began with crude (but for a time effective) efforts with keyword stuffing yielded to increasingly sophisticated tactics: clickbaiting, backlink purchasing, link swapping, and other efforts meant to allow mediocre or even bad content—from either a quality or relevance perspective, and often both—to perform better than it should have performed. As the algorithm changed, these SEO sharks would analyze it, test new methodologies against it, and apply it while churning out more of the same garbage.
The other SEO writers, whom we might consider to represent thArtificial Intelligencee light side of the force,focused not on getting a fix on the algorithm, but rather on creating the content that Google and Bing are trying to lift up. Badly written content and unreliable information have littered the internet landscape since its first years in the public eye, and the search algorithms are intended to help the best of what is out there, and the most relevant to particular users, rise to the top of results lists. SEO providers who work with Google and Bing, rather than against them, provide readers with what they want, rather than luring readers to sites that are less helpful.With search engine AI still in its relative infancy, we are already seeing that latter school of SEO strategy gaining a big advantage. The ability of AI tools to sort out what really matters to people searching has lessened the impact of gaming strategies and enhanced the ability to sort out and identify the sites that provide the requested information. On the other hand, there is still a long way to go; SEO providers who game the system can still target the other signals in the Hummingbird algorithm and artificially inflate search rankings through a number of techniques.In this way, the current state of AI matters, but is most important in what it means for the future of search as the tools develop.
How AI May Impact Web Design
When you look at how AI may deliver more on its promise in the future, you can begin with web design. Interestingly enough, SEO Web Designing already contains components of AI within itself. Responsive web design, for example, allows a site to adjust itself to the size and layout of the screen on which the site is being accessed. Similarly, techniques that site builders currently employ allow the frames and templates that a site employs to flex and contract as site owners add content. Where a frame built for large volumes of content can sometimes look absurd when only a few paragraphs are added, the site can now adjust the display so the frame and content fit perfectly together.
From the other end, as internet search AI tools develop, the search engines should be able to take into account site design issues in ways that they have only touched on up to this point.Images and videos currently rely in part on metatext and scripts to get noticed by search engines. As AI syncs up with optical recognition software, though, the opportunities abound for web designers to immerse a site in imagery, sound, and even borders that tie the site to the concepts discussed in the text. The search engines have already become much better in the last decade at this kind of connection; the best SEO providers will work to connect web design and SEO written content in ways that build powerfully together.
The reverse will also be true, much to the chagrin of click baiters and other purveyors of search shenanigans. The concept of providing lurid images, or promises of shocking information, and connecting it with boring or useless content should lose its impact in the coming years. The more sophisticated the tools become in linking search intent with website results, the less effective this kind of disjointed web design should become.
Social Media and Ads
Besides deceptive designs, links from social media accounts and paid advertisements will likely experience profoundly changed effectiveness in the years ahead. Already, Facebook is working to lessen the impact that paid advertising has on its news feeds for users. Advertisers are treating this as a blow to their efforts to reach high volumes of users for relatively low cost or energy; having long capitalized on the Social Network’s desire to build income streams, those posting will now need to adjust their strategies.
we can begin to assess what AI means to SEO work, both now and in the foreseeable future.Here, again, the ability to create interesting content and gain social media attention should provide the strongest content providers with significant market advantages. The power of Social Media Marketing lies with the exponential growth of a message’s reach. Rather than use payments and clickbait to entice readership, this approach will reward truly shareworthy material to rise above the rest. The lower dilution in this approach will mean stronger results for those who do it right.
We can expect a similar impact over time with Google ads and other online paid advertising. Limited shelf space is available to those creating and posting ads. AI tools will likely improve over time at positioning the best, most relevant ads in the places they will have the best impact. This will serve advertisers well, because higher conversion rates should follow. It will also serve the best SEO providers well, because if they connect PPC ads to powerful content, they should be rewarded accordingly.
Why SEO Fears Are Overblown
All in all, any harm in AI to the SEO providers will fall on those working hardest to not have to provide great content. It will mean many content providers and SEO designers need to change their focus, but this should prove to be a good thing to many people. The things that have always been intended to matter most will continue to do so, and the highest-quality, most relevant content should more effectively reach the people who want it.
But what of a future that might include AGI or ASI? These feel far-fetched; computers performing the same way the human mind performs would be incredible, because machines are not generally equipped to function outside of the parameters of logical flow and processing.A computer that understands in the way we do would represent a fundamental shift not just in technology, but in the basic functionality applied.
Still, many do feel they are possible. Truthfully, in examining the two, it is easier to envision ASI than AGI; computers performing at the same level as human minds, but with the processing speed and capabilities that they bring to the tasks on which they are employed, would quickly advance past the state of equality to superior performance. If they were to become capable of the kind of thinking that people do, they would almost immediately become capable of more.
Should this move from sci-fi concept to concrete fact, one could expect the impact on SEO to be immense. AI would become not a signal in an overall algorithm as it (arguably) is today, but rather replace the algorithm entirely. SEO would become a matter not of entering search terms so much as conducting an analytical conversation, and then bringing up incredibly precise results. Here, you would expect Google to go out of business before SEO providers, though.After all, the focus would remain—and even become more than ever—on providing excellent content that meets specific needs. Gamesmanship would all but disappear, and the best content providers would win the day.
Unless, of course, the ASI applications developed emotions and resented their lot in life. As the movies tell us, at this point SEO techniques would become the least of our concerns.