Artificial Intelligence is increasingly being used in SEO and an understanding of how this works ensures you are with the learning curve Google has introduced through Rankbrain, a new AI machine learning algorithm which is a game changer for how search results are processed.
There are currently three definitions for Artificial Intelligence which reflect it’s uses and developmental capabilities.
- Artificial Narrow Intelligence is when AI is used for one focussed, targeted outcome ie spam filters.
- Artificial General Intelligence describes AI that can process information as a human brain would, with an appreciation of changing context and complex variables.
- Lastly Artificial Super Intelligence is when AI is performing at a level beyond the human brain.
Pedro Domingos in the Master Algorithm identifies Google’s Rankbrain as based on thinking from the Connectionist school of thought. Connectionist’s work from the belief that knowledge is encoded through connection, similar to the neurons in our brains. Rankbrain is based on this belief process which is rebranded as ‘deep learning’.
This means the SEO environment is changing rapidly, loopholes are being closed and it is becoming increasingly important to be technically savvy to succeed in this new SEO environment.
Within Google there are a number of core algorithms. It is Rankbrain’s task to identify the optimal use and ordering of these for each search result. Each search result within Google becomes it’s own footprint for a unique mix of algorithms.
At the vertical level this changes how to analyse data over time where good practice becomes the new order of the day as dictated by Google. Therefore within your niche area it will become the trend to emulate the best examples of websites as a way of staying in the flow as Rankbrain learns what the correct settings are for each environment and promotes these ‘good’ sites and other sites as ‘bad’.
It becomes more problematic when you have a site that spans a range of industries or subject matter as this is more difficult for Rankbrain to respond to. So the prioritising of information in this case becomes more random with less predictable outcomes as the deep learning model struggles to classify sites of this nature. An example of this would be a site like the How to sites. It is worth following this trend to learn how this aspect of SEO and AI unfolds over time.
With regards to the future of backlinking, Rankbrain compares back linking profiles and identifies if something is ‘off’. Rankbrain classifies and responds to something as being off if it deviates from what it has identified as ‘good’. This means organic traffic and new leads are more challenging for it to classify as they don’t conform to a pattern or type. It will become increasingly important to remain within your linking neighbourhood if you want to be understood correctly.
Finding ways of protecting your organic flow of traffic from a deep learning process that currently struggles to recognise value on what is not classifiable in its own terms will be one of the challenges that AI can hopefully resolve in the future.