Google has developed a new technology aimed at using artificial intelligence (AI) to answer extremely complex search queries. Last month at its I/O 2021 conference, Google introduced MUM (Multitask United Model). According to Google, this language model is 1000x more powerful than BERT, the language model released in 2018 that is now used on Google searches in over 70 languages.
One of MUM’s biggest strengths is that it can process multiple tasks at the same time, such as reading text, understanding, finding audio and video to inform responses, and forming insights from more than 75 languages. It can also translate and compile those findings simultaneously to answer complex questions that require complex answers.
Google’s vision is to one day enable MUM to answer questions like: “I’ve hiked Mt. Adams and now want to hike Mt. Fuji next fall. What should I do differently to prepare?”
Today, users would have to create many queries for different pieces of information to answer that question. Users would need information like elevation of the mountains, average temperatures, trail difficulty, gear options, and more to know what to search for. Google says MUM would be able to deliver a similar response to one that a hiking expert would give to the same question, meaning it can answer questions you may not yet know you have, such as offering insight into unknown terrain or weather patterns.
Powered by Artificial Intelligence
While there’s currently not a lot of public information about how MUM functions, it’s likely that MUM will incorporate many of the AI projects Google has already been working on for years. There seems to be a lot of overlap between all of these AI advancements, from the conversational agents that can understand complex text inputs to breakthrough computer vision models that can do object detection and understand videos. MUM will combine these approaches in real time to help with better searches.
For consumers, MUM means Google will soon provide much more value in its search engine, and Google Search will be much more contextual. They’ll be able to get more information with fewer queries and rich results that add to text answers. Google’s goal is to deliver responses that are personalized based on your needs and that can also understand the same nuances as humans. MUM is not the final evolution of Google’s quest to optimize their search results, but it goes a long way to helping users get the info they want.
Why Brands Should Care
Marketers will also benefit from these updates, as it’ll ensure more relevant content is being delivered to consumers at the right time. Brands should be thinking about creating content that provides value to consumers and can be used in more contextually relevant responses. As always, brands will benefit from having a plan, and one that creates content for consumers at every stage of their journey. Including rich media and giving Google more options to pull from allows for more contextual responses.
Google has said that it won’t be mum on MUM, and we’ll know as soon as it’s added to searches, as with BERT. We’ll continue to monitor how Google will implement MUM into search and other products. For more information about AI or optimizing your brand for search, reach out to your Intouch team.