It is official: there is no way to avoid the influences of artificial intelligence in search results. Google built a machine-learning AI called RankBrain. Its use was confirmed in 2015. The AI’s reach was initially limited to just 15% of searches. Now, in 2018, RankBrain is handling all queries.
How Does Machine-Learning AI Work?
When RankBrain comes across a phrase or word that it doesn’t know, it will make a guess based on terms it knows that could have a similar meaning. The search results are filtered accordingly. This guess is recorded and adapted over time to give the user better results. It sounds like a hit-or-miss system, but it is very effective.
Google has been experimenting with artificial intelligence for years. In 2017, researchers announced the development of AutoML, an AI that could generate another AI that outperformed those made by humans.
Optimizing for Humans Is Still Recommended
There are still issues that come with the use of AI in search engine results. It may present a “good enough” solution that isn’t as refined is it could be. AI can also confuse correlation with causation which can form connections between unrelated content. Developers program the machine-learning algorithm, which means they don’t fully understand the nuances of the final algorithm. That makes it difficult to predict how search results will behave in the future.
Optimizing for humans is still recommended. The goal of AI is to make it think like a person. As the user and search engine are more closely matched, producing content for the former will also benefit a site’s standing with the latter. Find out more about search engine SEO by contacting OMA Comp.