How Does Google Use AI in Search?

As Google continues to benefit from more AI (Artificial Intelligence) and machine learning in Google Search, one might wonder how AI and machine learning help that Google Search performs its daily tasks. Since 2015, when Google introduced its first AI called RankBrain, Google has continued to deploy AI systems to understand better the language and thus improve the search results which Google delivers in its searches.

A few months ago, we sent Google a series of questions about how Google uses its AI in search, including RankBrain, neural mapping, BERT and MUM which Google's latest AI breakthrough. We got the information more about when Google uses AI, which AI does what in Google Search how these various AI algorithms might work together, how they have changed over the years and what search marketers need. Learn how Google uses artificial intelligence in search.

We talked to Google Search PR Officer Danny Sullivan to help us answer many of these questions. In short, RankBrain, neural mapping and BERT are used in many ways queries in Google's ranking system, trying to understand the language of both the query and the content. Although, MUM is not currently used for sequencing purposes, it is currently only used for COVID vaccine naming and powers related topics in the video results.

It starts by writing content for people

"Write content for people" is a sentence that you hear all the time from Google reps and many SEO experts. In the old days of SEO, when algorithms were perhaps simpler, you would have many SEOs to create content for each search engine. (There were dozens of different search engines back then) Now, there is Google primarily combination with Bing and a bit of DuckDuckGo - but algorithms are much more complex and with machine learning and AI, algorithms understand the language as the same way a human would.

So the advice Google gives is to write for people and you cannot optimize your site for BERT or any AI. If you write content that people understand, the algorithms and AI search engines will understand it, too. In short, this article is not intended to give you SEO tips on how to optimize your sites for any AI, but rather how Google uses AI in Google Search.

Overview of artificial intelligence used in Google Search

RankBrain. Google's first attempt to use AI in search begins with RankBrain, which dates back to 2015. Google said that RankBrain helps Google understand how words relate to concepts and can take a broad query and better define how that query relates to real-world concepts. Launched in 2015 and used in 15% of queries, Google has said that it is now widely used in many queries and in all languages and regions by 2022. RankBrain specifically helps Google to rank search results and is part of ranking algorithm.

- Launch Year: 2015
- Used for Sorting: Yes
- Looks at query and content language
- Works for all languages
- Very widely used for many queries

Here is an example of how RankBrain is used by Google, if you search "what is the title of the consumer at the top of a food chain", Google systems will see these words on various pages and learn what a food concept means. This chain can be related to both animals and humans. By understanding these words and matching them with their related concepts, RankBrain helps Google understand what you are looking for in what is commonly referred to as an "apex predator".

Neural mapping. Neural mapping was the next AI of Google released for search, it was released in 2018 and then expanded to local search results in 2019. Actually, we have an article here explaining the differences between RankBrain and neural mapping. Google told us that neural mapping helps Google understand how queries relate to pages by looking at the entire query or content on the page and understanding it in the context of that page or query. Today, neural mapping is used in most queries, all languages, all regions and most search verticals, if not most searches. Neural mapping specifically helps Google to rank search results and is part of the ranking algorithm.

- Launch Year: 2018
- Used for Sorting: Yes
- Looks at query and content language
- works for all languages
- Very widely used for many queries

For example, if you search for "insights on how to manage green," an example of how neural mapping is used by Google is provided. “If a friend asks you this, you'd probably be surprised,” Google said. Google said that "However, with neural matching, we are able to make sense of this interrogative search. By looking at broader representations of the concepts in the query (management, leadership, personality and more), neural mapping can decipher this researcher's search for management clues based on a popular, color-based personality guideline."

BERT. BERT, Bidirectional Encoder Representations from Transformers, is a neural network-based technique for natural language processing pre-training in 2019. Google told us that BERT helps Google understand how word combinations convey different meanings and purposes, including looking at the word order on a page, so even seemingly unimportant words in your queries are counted. When BERT was introduced, it was used in 10% of all English queries but it was expanded to more languages and was used early in almost in all languages. BERT specifically helps Google to rank search results and is part of ranking algorithm.

- Launch Year: 2019
- Used for Sorting: Yes
- Looks at query and content language
- Works for all languages, but Google said BERT "plays a critical role in almost every English query."
- Very widely used for many queries

A Google-provided example of how BERT is used helps us understand that if you search "Can you buy medicine for someone pharmacy?"BERT helps us understand whether you buy medicine or not. Before BERT, we took this short preposition to be precise, mostly figuring out how to fill a prescription. Google said.

MUM. MUM the Multitask Unified Model, is Google's newest artificial intelligence in search. MUM was introduced in 2021 and then expanded again at the end of 2021 for more applications with many promising uses in the future. Google told us that MUM can be used to not only understand languages, but also new terms and understanding variations in languages. MUM is not currently used for any ranking purposes in Google Search, but supports all languages and regions.

- Launch Year: 2021
- Used for Sorting: No
- Not query or language specific
- Works for all languages, but Google is not used for ranking purposes today.
- Used for a limited number of purposes

Currently, MUM is used to improve searches for COVID-19 vaccine information, and Google said it is “looking forward to offering more intuitive ways to search using a combination of both text and images in Google Lens in the coming months.”

AI is used together in search but can be customized for search sectors

Danny Sullivan of Google also explained that while these are individual AI-based algorithms, they often work together to help sort and understand the same query.

Google said that all these AI systems are "used to understand the language, including the query and potentially relevant results" and are "not designed to act individually to just analyze a query or a page." Previously, it may have assumed and understood that an AI system might look more at understanding the query and not the content on the page, but that is not the case, at least not in 2022.

Google also confirmed in 2022 that RankBrain, neural mapping and BERT are used globally in all languages which Google Search runs in.

And when it comes to web search and local search, as well as images, shopping and other industries, Google has announced that it uses RankBrain, neural mapping and BERT for web Search. According to Google, other modes or sectors of Google Search, such as images or shopping mode, use separate proprietary AI systems.

What about core updates and artificial intelligence?

As explained above, Google uses RankBrain, neural mapping and BERT on most queries you enter into Google Search, but Google has core updates as well. The extensive Google updated that Google delivers several times a year are often more noticeable to site owners, publishers and SEOs than when Google introduced these larger AI-based systems.

However, Google said that all of these could work with core updates. Google said that RankBrain, neural mapping and BERT are larger AI systems they have. However, they have many AI systems in search and some of the core updates Google offers.

Google said they have other machine learning systems on Google Search. "RankBrain, neural mapping and BERT are just a few of our more powerful and prominent systems." Google said. Google also added that "There are other AI elements that may affect core updates that do not belong to these three specific AI systems."

Source: searchengineland.com