Google's Search Engine May Start Running On AI, But Not Anytime Soon

A comment from John Mueller highlighted the use of machine learning in Google Search's development, and how much bearing it'll have on the future.

John Mueller, holding the position of Senior Webmaster Trends Analyst, has been working at Google since 2007. With that long-term reputation in his tool-belt, combined with how vocal the man is on Twitter, Mueller is a well-known and trusted font of knowledge regarding all things Google.

Machine learning and AI is quickly being looked at as the face of our future. With our fast-paced generation, tech companies are looking to automate performance and functionality as much as possible. Removing the human element goes a long way in making efforts more time-saving and cost effective. This does, obviously, come with its issues. Machine learning, while useful, is a process that takes time to develop and even more time to implement effectively. Due to AI not really having any sort of sentience, machine learning based programs can't make the snap decisions of judgement in unprecedented situations that humans can. And Google seems to be very aware of this.

Bing, the Microsoft owned search engine and a competitor of Google Search, has been revealed to run almost entirely on AI/machine learning. The revelation, courtesy of an interview panel with Bing's representatives during the SMX West convention, that 90% of the engine runs by artificial means is certainly interesting, and paints a fun image of the running algorithm.

Google representatives have always been tight-lipped about delving into their algorithms and how those are run. During that very same convention, Google reps stated that machine learning will not take over the Search website's ranking algorithm (yet). While it features some amount of AI (no specifics were provided, short of the usage of RankBrain), the keyword rankings still heavily rely on classic algorithm design. The involvement of AI and machine learning would also warrant major amounts of debugging, which is an arduous and long-term process.

SMX West was held in early 2019. Now, coming at the tail end of 2020, it seems that while Google seems to be expanding more into machine learning, the approach being taken is careful. John Mueller states how some algorithms are more suitable than others. He also echoes the same concern of debugging, except worded differently. The entire tweet reads as if Google's fine with the implementation of AI, so long as its provided adequate time and care.

AI may be the face of the future, but its still experimental technology. While Google is certainly looking to place a firm foot into the future, it certainly is taking much time and care in order to ensure that the transition is smooth, seamless, and causes no hindrance to the billions that rely on Google Search.



Read next: Make Al Look More “Positive” says Google

No comments:

Post a Comment