- Google has amped up its search engine and its capacities by introducing a new Artificial Intelligence-based algorithm.
- This new search engine model will allow users to receive multimodal information for their queries by simultaneously providing answers via text, image as well as videos.
- The utility of this algorithm is still under the scanner as research has been published about the racial and gender bias this machine learning platform can induce.
On 29th September, Google launched its Search On event during which it introduced revolutionary new changes. These features appear as the company’s most radical attempt at enhancing its search engine. Powered with these new tools users can expect to receive in-depth context-rich answers for their questions as entered in the search box without unnecessary complications of ads and unuseful data.
What is Google’s new algorithm?
Like several technological giants of the current era, Google has decided to shift to a greater Artificial Intelligence(AI) dominated algorithm. The new Multitask Unified Model(MUM) machine learning technology was unveiled as a part of the Search On event. This comes as the latest addition to Google Search Algorithm which is based on AI and exploits its abilities to understand the search engine user’s query on multimodal platforms. That means this new algorithm will analyze different forms of information like text, image, and video simultaneously to give combined results.
Armed with the new MUM algorithm, users will witness a new search engine interface. The new redesign will begin by showing the user the “things to know” regarding their search. This will be formatted as boxes on clicking which, users will be directed towards websites catering information on the various subtopics of the original question.
Videos on the web containing a mention of one’s query will be listed according to the contextual relevance. Shopping-related queries will give detailed results regarding the various websites said product is available at while also comparing prices simultaneously. Such queries will also list out local stores where said products are available.
Google is also pushing forward to include more apps to use its image recognition software called Google Lens. Google plans to stop marketing Google Lens as a separate app and wants to make it appear as an inbuilt option in the Google app as well as the Chrome browser for both Android and iOS platforms. The lens will allow users to take a picture of their query, suppose an automobile part they wish to purchase, a text in a foreign language to be translated, or simply a flower they wish to identify and provide exhaustive information regarding the same through recognition of the query and following local and international solutions. Powered with the machine learning MUM algorithm, the search will become more accurate, more explorative, and more useful.
What will change from the user’s perspective?
The MUM algorithm requires users to respect its multimodal approach by providing similar inputs. In simple terms, to enjoy Google’s powerful new AI-based search engine and its abilities to provide context-rich solutions, users need to input their queries on multiple formats including text, image, and video. This might appear again as Google’s attempt to forward the use of Google Lens as it is indeed a powerful tool.
However, users needn’t worry too much as MUM has gained popularity as a machine learning model which means, that the Ai model will monitor and try to understand your patterns. These patterns may be as simple as common websites you buy from or visit more advanced setups where the patterns of your users are recorded and divided into the number of times and duration. Either way, users can expect a powerful directed search experience.
Like every brand of its category, one of the most crucial setbacks observed in Google’s new search engine is the bias induced by the MUM model. When users type in their query in the search box, it’s observed that Google intentionally redirects it to websites promoting the company’s very own services or products. Other than this, another dangerous problem associated with Google’s new model is its inclination towards racial and gender bias.
Since this model is based on a machine learning interface, it is so designed to read large portions of exchanged information that circulates the web. Hence, it comes into contact which such content which may unintentionally tip off the engine to produce results for its users with said bias. This issue of bias has been researched thoroughly and published by Google’s very own lead researchers of the AI ethics team which led them to be fired. How useful or detrimental this technology will prove to be will be interesting to note.